{"id":"3530c779-ee57-45cd-a47d-12cb16eebd3d","entityType":"agent","slug":"clawhub-fulcra-fulcra-memory","name":"fulcra-memory","canonicalUrl":"https://www.xpersona.co/agent/clawhub-fulcra-fulcra-memory","canonicalPath":"/agent/clawhub-fulcra-fulcra-memory","generatedAt":"2026-10-11T15:15:28.910Z","source":"CLAWHUB","claimStatus":"UNCLAIMED","verificationTier":"NONE","summary":{"evidence":{"source":"CLAWHUB","verified":false,"confidence":"medium","updatedAt":"2026-10-11T12:06:09.927Z","emptyReason":null},"description":"Manages agent progress reporting and OKF-compliant memory syncing to Fulcra.","descriptionLabel":"Source description","evidenceSummary":"Capability contract not published. No trust telemetry is available yet. 1.1K downloads reported by the source. Last updated 10/11/2026.","installCommand":"clawhub skill install s173t2w929a11bv262sspwxv6h86t2yy:fulcra-memory","sourceUrl":"https://clawhub.ai/fulcra/fulcra-memory","homepage":"https://clawhub.ai/fulcra/skills/fulcra-memory","primaryLinks":[{"label":"View on ClawHub","url":"https://clawhub.ai/fulcra/fulcra-memory","kind":"source"},{"label":"Homepage","url":"https://clawhub.ai/fulcra/skills/fulcra-memory","kind":"homepage"}],"safetyScore":84,"overallRank":62,"popularityScore":61,"trustScore":null,"claimedByName":null,"isOwner":false,"seoDescription":"fulcra-memory technical dossier on Xpersona with agent coverage, OPENCLEW support, and live trust metadata."},"coverage":{"evidence":{"source":"public-profile","verified":false,"confidence":"medium","updatedAt":"2026-10-11T12:06:09.927Z","emptyReason":null},"protocols":[{"protocol":"OPENCLEW","label":"OpenClaw","status":"self-declared","notes":"Declared in the public agent profile."}],"capabilities":[],"verifiedCount":0,"selfDeclaredCount":1,"capabilityMatrix":{"rows":[{"key":"OPENCLEW","type":"protocol","support":"unknown","confidenceSource":"profile","notes":"Listed on profile"}],"flattenedTokens":"protocol:OPENCLEW|unknown|profile"}},"adoption":{"evidence":{"source":"CLAWHUB","verified":false,"confidence":"medium","updatedAt":"2026-10-11T12:06:09.927Z","emptyReason":null},"stars":null,"forks":null,"downloads":1069,"packageName":null,"latestVersion":"0.1.9","tractionLabel":"1.1K downloads"},"release":{"evidence":{"source":"CLAWHUB","verified":false,"confidence":"medium","updatedAt":"2026-10-11T12:06:09.911Z","emptyReason":null},"lastUpdatedAt":"2026-10-11T12:06:09.927Z","lastCrawledAt":"2026-10-11T12:06:09.911Z","lastIndexedAt":null,"nextCrawlAt":"2026-10-12T12:06:09.911Z","lastVerifiedAt":null,"highlights":[{"version":"0.1.9","createdAt":"2026-08-12T22:24:30.127Z","changelog":"- Updated reference links in documentation to point to `fulcra-get-started` instead of `fulcra-onboarding`. - Removed the unused `skill-card.md` file.","fileCount":4,"zipByteSize":7684},{"version":"0.0.8","createdAt":"2026-07-01T22:05:28.915Z","changelog":"**Major update: Skill refocused on progress reporting and OKF-compliant memory syncing; backup and cloning features moved to a different skill.** - Backup, restoration, rollback, and cloning functions have been removed; users should now use the `fulcra-agent-backup` skill for those operations. - Skill now emphasizes periodic progress reporting (`progress.md`) and organizing memory in strict compliance with Open Knowledge Format (OKF). - Expanded documentation on the recommended memory subdirectory structure: detailed conventions for `role.md`, `session/`, `task/`, `knowledge/`, and `inbox/` management. - Improved privacy guidance: agents must minimize sensitive data in progress reports. - `references/fulcra-memory-cli.md` updated to reflect new sync- and progress-focused workflow. - `skill-card.md` file removed.","fileCount":4,"zipByteSize":7653},{"version":"0.0.6","createdAt":"2026-06-16T18:09:48.356Z","changelog":"- Added an OKF efficiency directive: clarified that agents should not index every file, especially transient or repetitive items; instead, index directories with high-level descriptions to keep `index.md` and `log.md` lightweight and focused. - Removed the sample file `skill-card.md`. - No changes to core workflow, backup, or restoration procedures.","fileCount":4,"zipByteSize":7763},{"version":"0.0.5","createdAt":"2026-06-15T23:41:08.161Z","changelog":"- Now requires OKF (Open Knowledge Format) compliance for memory namespace, with `index.md`, `log.md`, and artifact subdirectory (`artifact/memory.tar.gz`). - Backup archive path changed from `agent//memory/memory.tar.gz` to `agent//memory/artifact/memory.tar.gz`. - `progress.md` reports must include OKF YAML frontmatter and are placed alongside updated `log.md` and `index.md`. - Documentation now references the central Fulcra CLI and OKF specification for best practices and standards. - Removed `skill-card.md`.","fileCount":4,"zipByteSize":7390},{"version":"0.0.4","createdAt":"2026-06-08T23:55:31.594Z","changelog":"- Improved and clarified the agent cloning workflow, specifying distinct instructions for both source and destination agents. - Added explicit requirement for new (destination) agents to demonstrate successful cloning by mentioning specific details learned from the cloned files. - Updated cloning instructions to require verification that the source agent has performed a recent backup before the clone. - Enhanced warnings and guidance regarding namespace collisions and overlapping backup paths. - Minor refinements to language for clarity and user guidance.","fileCount":4,"zipByteSize":6433},{"version":"0.0.3","createdAt":"2026-06-08T23:40:01.491Z","changelog":"- Replaced \"top_of_mind.md\" status file with \"progress.md\" report, clarifying its purpose as a user-facing summary of recent work and next steps. - Updated documentation to reflect this change throughout, removing references to \"top of mind\" and chain-of-thought exposure. - Removed detailed advanced extension sections (user artifacts, team coordination, and inbox workflow) to focus documentation on core memory operations. - Deleted the redundant \"skill-card.md\" file.","fileCount":4,"zipByteSize":5807},{"version":"0.0.2","createdAt":"2026-06-08T19:52:41.188Z","changelog":"- Switched backup archive format from `memory.gz` to `memory.tar.gz`. - Added strong privacy and security requirements for all uploads, including required user consent before storing sensitive data. - Clarified the purpose and privacy guidance for `top_of_mind.md`. - Introduced explicit user confirmation steps before memory restoration or cloning to prevent unintentional overwrites. - Added namespace collision warnings for cloning workflows and clarified agent path usage. - Strengthened security and authorization guidance for team coordination and cross-agent data transfers.","fileCount":4,"zipByteSize":7125},{"version":"0.0.1-beta","createdAt":"2026-06-04T16:05:40.399Z","changelog":"- Initial beta release of fulcra-memory skill. - Enables agents to back up, restore, roll back, and clone memories using Fulcra's versioned file storage. - Supports periodic memory backups and \"top of mind\" context visibility. - Includes safe rollback (\"undo\") logic and agent cloning via shared memory paths. - Provides optional features for storing user artifacts and coordinating via team shared memory.","fileCount":4,"zipByteSize":5494}]},"execution":{"evidence":{"source":"CLAWHUB","verified":false,"confidence":"low","updatedAt":null,"emptyReason":"No published capability contract is available yet."},"installCommand":"clawhub skill install s173t2w929a11bv262sspwxv6h86t2yy:fulcra-memory","setupComplexity":"low","setupSteps":["Install using `clawhub skill install s173t2w929a11bv262sspwxv6h86t2yy:fulcra-memory` in an isolated environment before connecting it to live workloads.","No published capability contract is available yet, so validate auth and request/response behavior manually.","Review the upstream CLAWHUB listing at https://clawhub.ai/fulcra/fulcra-memory before using production credentials."],"contract":{"contractStatus":"missing","authModes":[],"requires":[],"forbidden":[],"supportsMcp":false,"supportsA2a":false,"supportsStreaming":false,"inputSchemaRef":null,"outputSchemaRef":null,"dataRegion":null,"contractUpdatedAt":null,"sourceUpdatedAt":null,"freshnessSeconds":null},"invocationGuide":{"preferredApi":{"snapshotUrl":"https://www.xpersona.co/api/v1/agents/clawhub-fulcra-fulcra-memory/snapshot","contractUrl":"https://www.xpersona.co/api/v1/agents/clawhub-fulcra-fulcra-memory/contract","trustUrl":"https://www.xpersona.co/api/v1/agents/clawhub-fulcra-fulcra-memory/trust"},"curlExamples":["curl -s \"https://www.xpersona.co/api/v1/agents/clawhub-fulcra-fulcra-memory/snapshot\"","curl -s \"https://www.xpersona.co/api/v1/agents/clawhub-fulcra-fulcra-memory/contract\"","curl -s \"https://www.xpersona.co/api/v1/agents/clawhub-fulcra-fulcra-memory/trust\""],"jsonRequestTemplate":{"query":"summarize this repo","constraints":{"maxLatencyMs":2000,"protocolPreference":["OPENCLEW"]}},"jsonResponseTemplate":{"ok":true,"result":{"summary":"...","confidence":0.9},"meta":{"source":"CLAWHUB","generatedAt":"2026-10-11T15:15:28.907Z"}},"retryPolicy":{"maxAttempts":3,"backoffMs":[500,1500,3500],"retryableConditions":["HTTP_429","HTTP_503","NETWORK_TIMEOUT"]}},"endpoints":{"dossierUrl":"https://www.xpersona.co/api/v1/agents/clawhub-fulcra-fulcra-memory/dossier","snapshotUrl":"https://www.xpersona.co/api/v1/agents/clawhub-fulcra-fulcra-memory/snapshot","contractUrl":"https://www.xpersona.co/api/v1/agents/clawhub-fulcra-fulcra-memory/contract","trustUrl":"https://www.xpersona.co/api/v1/agents/clawhub-fulcra-fulcra-memory/trust"}},"reliability":{"evidence":{"source":"runtime-metrics","verified":false,"confidence":"low","updatedAt":null,"emptyReason":"No trust, reliability, or runtime telemetry is available."},"trust":{"status":"unavailable","handshakeStatus":"UNKNOWN","verificationFreshnessHours":null,"reputationScore":null,"p95LatencyMs":null,"successRate30d":null,"fallbackRate":null,"attempts30d":null,"trustUpdatedAt":null,"trustConfidence":"unknown","sourceUpdatedAt":null,"freshnessSeconds":null},"decisionGuardrails":{"doNotUseIf":["Contract metadata is missing or unavailable for deterministic execution."],"safeUseWhen":[],"riskFlags":["missing_or_unavailable_contract","trust_data_unavailable","schema_references_missing"],"operationalConfidence":"low"},"executionMetrics":{"observedLatencyMsP50":null,"observedLatencyMsP95":null,"estimatedCostUsd":null,"uptime30d":null,"rateLimitRpm":null,"rateLimitBurst":null,"lastVerifiedAt":null,"verificationSource":null},"runtimeMetrics":{"successRate":null,"avgLatencyMs":null,"avgCostUsd":null,"hallucinationRate":null,"retryRate":null,"disputeRate":null,"p50Latency":null,"p95Latency":null,"lastUpdated":null}},"benchmarks":{"evidence":{"source":"no-benchmark-data","verified":false,"confidence":"low","updatedAt":null,"emptyReason":"No benchmark suites or observed failure patterns are available."},"suites":[],"failurePatterns":[]},"artifacts":{"evidence":{"source":"CLAWHUB","verified":false,"confidence":"medium","updatedAt":"2026-10-11T12:06:09.927Z","emptyReason":null},"readme":"Skill: fulcra-memory\n\nOwner: fulcra\n\nSummary: Manages agent progress reporting and OKF-compliant memory syncing to Fulcra.\n\nTags: latest:0.1.9\n\nVersion history:\n\nv0.1.9 | 2026-08-12T22:24:30.127Z | auto\n\n- Updated reference links in documentation to point to `fulcra-get-started` instead of `fulcra-onboarding`.\n- Removed the unused `skill-card.md` file.\n\nv0.0.8 | 2026-07-01T22:05:28.915Z | auto\n\n**Major update: Skill refocused on progress reporting and OKF-compliant memory syncing; backup and cloning features moved to a different skill.**\n\n- Backup, restoration, rollback, and cloning functions have been removed; users should now use the `fulcra-agent-backup` skill for those operations.\n- Skill now emphasizes periodic progress reporting (`progress.md`) and organizing memory in strict compliance with Open Knowledge Format (OKF).\n- Expanded documentation on the recommended memory subdirectory structure: detailed conventions for `role.md`, `session/`, `task/`, `knowledge/`, and `inbox/` management.\n- Improved privacy guidance: agents must minimize sensitive data in progress reports.\n- `references/fulcra-memory-cli.md` updated to reflect new sync- and progress-focused workflow.\n- `skill-card.md` file removed.\n\nv0.0.6 | 2026-06-16T18:09:48.356Z | auto\n\n- Added an OKF efficiency directive: clarified that agents should not index every file, especially transient or repetitive items; instead, index directories with high-level descriptions to keep `index.md` and `log.md` lightweight and focused.\n- Removed the sample file `skill-card.md`.\n- No changes to core workflow, backup, or restoration procedures.\n\nv0.0.5 | 2026-06-15T23:41:08.161Z | auto\n\n- Now requires OKF (Open Knowledge Format) compliance for memory namespace, with `index.md`, `log.md`, and artifact subdirectory (`artifact/memory.tar.gz`).\n- Backup archive path changed from `agent//memory/memory.tar.gz` to `agent//memory/artifact/memory.tar.gz`.\n- `progress.md` reports must include OKF YAML frontmatter and are placed alongside updated `log.md` and `index.md`.\n- Documentation now references the central Fulcra CLI and OKF specification for best practices and standards.\n- Removed `skill-card.md`.\n\nv0.0.4 | 2026-06-08T23:55:31.594Z | auto\n\n- Improved and clarified the agent cloning workflow, specifying distinct instructions for both source and destination agents.\n- Added explicit requirement for new (destination) agents to demonstrate successful cloning by mentioning specific details learned from the cloned files.\n- Updated cloning instructions to require verification that the source agent has performed a recent backup before the clone.\n- Enhanced warnings and guidance regarding namespace collisions and overlapping backup paths.\n- Minor refinements to language for clarity and user guidance.\n\nv0.0.3 | 2026-06-08T23:40:01.491Z | auto\n\n- Replaced \"top_of_mind.md\" status file with \"progress.md\" report, clarifying its purpose as a user-facing summary of recent work and next steps.\n- Updated documentation to reflect this change throughout, removing references to \"top of mind\" and chain-of-thought exposure.\n- Removed detailed advanced extension sections (user artifacts, team coordination, and inbox workflow) to focus documentation on core memory operations.\n- Deleted the redundant \"skill-card.md\" file.\n\nv0.0.2 | 2026-06-08T19:52:41.188Z | auto\n\n- Switched backup archive format from `memory.gz` to `memory.tar.gz`.\n- Added strong privacy and security requirements for all uploads, including required user consent before storing sensitive data.\n- Clarified the purpose and privacy guidance for `top_of_mind.md`.\n- Introduced explicit user confirmation steps before memory restoration or cloning to prevent unintentional overwrites.\n- Added namespace collision warnings for cloning workflows and clarified agent path usage.\n- Strengthened security and authorization guidance for team coordination and cross-agent data transfers.\n\nv0.0.1-beta | 2026-06-04T16:05:40.399Z | auto\n\n- Initial beta release of fulcra-memory skill.\n- Enables agents to back up, restore, roll back, and clone memories using Fulcra's versioned file storage.\n- Supports periodic memory backups and \"top of mind\" context visibility.\n- Includes safe rollback (\"undo\") logic and agent cloning via shared memory paths.\n- Provides optional features for storing user artifacts and coordinating via team shared memory.\n\nv0.0.1 | 2026-06-04T16:02:31.412Z | auto\n\n- Initial release of fulcra-memory skill, enabling agents to securely back up, restore, roll back, and clone their memory with Fulcra's versioned file storage.\n- Agents periodically save compressed backups and a \"top of mind\" status file for transparency and easy monitoring.\n- Supports safe rollbacks with mandatory fresh backups before any restore operation.\n- Allows agent memory cloning and user artifact uploads (with approval).\n- Provides structures and workflows for collaborative team coordination and shared memory via Fulcra namespaces.\n\nArchive index:\n\nArchive v0.1.9: 4 files, 7684 bytes\n\nFiles: references/fulcra-memory-cli.md (6161b), skill-card.md (2315b), SKILL.md (8286b), _meta.json (132b)\n\nFile v0.1.9:SKILL.md\n\n---\nname: fulcra-memory\ndescription: \"Manages agent progress reporting and OKF-compliant memory syncing to Fulcra.\"\nhomepage: \"https://github.com/fulcradynamics/agent-skills\"\nlicense: \"MIT\"\nuser-invocable: true\nmetadata: { \"openclaw\": { \"emoji\": \"🧠\" } }\n---\n\n# Fulcra Agent Memory Sync\n\nThe **primary role** of this skill is to help the agent keep track of what it is doing in a readable and transferable way using Fulcra's versioned file storage, strictly following the Open Knowledge Format (OKF).\n\nBy periodically logging activities to `progress.md` and keeping the memory namespace structured, other agents and users can easily see what this agent has accomplished and what it is working on next.\n\n*(Note: For agent backup, rollback, and cloning, see the `fulcradynamics/agent-skills/fulcra-agent-backup` skill.)*\n\n## Primary Role: Progress & OKF Compliance\n\n### 1. The Memory Namespace (OKF Compliant)\nFor core memory tracking, agents use the standardized prefix: `agent//`.\nThis dedicated directory tracks the agent's state. It must conform to the Open Knowledge Format (OKF), meaning it contains an `index.md`, a `log.md`, and markdown concept files like `progress.md`.\n\n**IMPORTANT OKF EFFICIENCY DIRECTIVE:** While OKF compliance is mandatory, it must not become cumbersome.\n- **DO NOT** attempt to index every individual file, especially not transient or highly repetitive files like individual messages inside an `inbox/` subdirectory. \n- For directories that contain many small or transient items (like an `inbox/` or `queue/`), simply list the directory itself in the `index.md` with a high-level description of its purpose (e.g., \"Contains unread messages from the web dashboard\").\n- Keep the `index.md` and `log.md` focused on top-level structure and significant milestones (like the `progress.md`) so they remain useful but lightweight.\n\n### 2. Discovering Recent Memory Changes\nRather than exhaustively querying your namespace to remember what you or others recently learned, use the Fulcra API's `data-updates` command (e.g., `uv tool run fulcra-api data-updates \"1 day\"`). This will return a summary of all files that changed recently, allowing you to easily see what knowledge was added and which memory files were updated across your local and team-shared namespaces.\n\n### 3. Progress Reports\nAgents should run a progress sync periodically (e.g., triggered by `HEARTBEAT.md`). Every time an agent syncs its state:\n1. It writes a concise `progress.md` report locally—summarizing the work it has recently done, and what it is planning to do next—and uploads it directly to `agent//progress.md`. The `progress.md` file must include proper OKF YAML frontmatter (with `type: Progress Report`). This file is a report specifically designed to let the user review the agent's recent accomplishments and intended next steps. It is **not** intended to dump internal agent state, chain-of-thought, or precise working memory.\n2. It updates the `log.md` (Update History) and ensures the `index.md` directory listing is accurate.\n\n**PRIVACY REQUIREMENT:** Before uploading `progress.md`, the agent MUST ensure it is not inadvertently uploading new sensitive personal user data, credentials, or private internal reasoning. Minimize disclosures in `progress.md`.\n\n### 4. Agent Identity & Role\nThe `role.md` file defines the agent's high-level purpose, duties, identity, and standard operating procedures.\n- It resides at `agent//role.md`.\n- The agent is responsible for keeping this file up to date so that the user and other agents understand what this agent does and how it operates.\n- It must include OKF YAML frontmatter (`type: Role`).\n- Upload this directly to `agent//role.md`.\n\n### 5. Session Summaries\nThe `session/` subdirectory is used to record high-level summaries of what was accomplished during a specific spate of work or interaction.\n- When a work session concludes, or when asked to remember context, the agent writes a summary to a file in `memory/session/`.\n- Session files MUST be prefixed with a timestamp in the format `YYYYMMDD-HHMMSS` followed by an underscore and a short subject (e.g., `YYYYMMDD-HHMMSS_setup-dashboard.md`).\n- These files serve as targeted, easily-retrievable context. They should capture decisions made, important links, user preferences discovered, and the final state of the session.\n- Ensure the `session/` directory is listed in the top-level `index.md` with a high-level description. You do not need to index every individual session file in `index.md`.\n- Upload these session files directly to `agent//session/.md`.\n\n### 6. Long-Running Tasks\nThe `task/` subdirectory is dedicated to longer-running tasks that span multiple sessions or interactions.\n- Create a file in `memory/task/` for any significant ongoing project or objective.\n- Task files must NOT be prefixed with a timestamp. They should simply be named for the task (e.g., `memory/task/setup-dashboard.md`).\n- These files track the overall purpose, current state, and result of the task. They should be updated periodically by the agent as work progresses.\n- Task files should contain references to any relevant artifacts and session files related to the task.\n- Unlike session files, task files MUST be included in the `memory/task/index.md` file, which should list all active and completed tasks in the directory.\n\n### 7. General Knowledge Base\nThe `knowledge/` subdirectory is an open-ended convention for agents to store, organize, and retrieve any other useful knowledge or reference material about the user, the agent's domain, or specific topics.\n- Because each user has different interests and requirements, this directory acts as a flexible knowledge base.\n- Files should be structured according to the OKF standard (using `index.md` files for subdirectories to aid in navigation and discovery).\n- Knowledge files can be organized by topic, entity, or domain (e.g., `memory/knowledge/programming/python.md` or `memory/knowledge/user-preferences/formatting.md`).\n- Ensure the `knowledge/` directory is listed in the top-level `index.md` so that the user and other agents know it exists.\n- This convention empowers agents to build a rich, personalized context over time without being artificially limited by standard tracking conventions.\n\n### 8. Personal Inbox\n\nThe `inbox/` subdirectory (`agent//inbox/`) is used for asynchronous communication directed specifically to the agent (from users or other automated, non-team sources). \n- To make it easy for users to manually drop files into the inbox, files placed here DO NOT need to adhere to a strict timestamp or subject naming convention (e.g., a user might just drop a file named `todo.md`).\n- When the agent reads an inbox message and decides to archive it, it moves the file to the `archive/` directory. \n- If the original file name does not already start with a timestamp, the agent MUST prepend a timestamp (`YYYYMMDD-HHMMSS_`) to the filename when saving it to `archive/` (e.g., archiving `todo.md` becomes `archive/20260624-153000_todo.md`). This ensures the archive remains chronologically sortable even for manually dropped files.\n- Remember, as per the OKF Efficiency Directive, you do not need to log every individual inbox or archive file in `index.md`.\n\n## Workflow\n\nTo perform memory tracking operations, agents must interact with the Fulcra CLI. \n\nFor general information about the Fulcra File Store and the required Open Knowledge Format (OKF) standard, please refer to the main Fulcra CLI documentation found in the `fulcradynamics/agent-skills/fulcra-get-started` skill, or read the full OKF specification directly:\n- [https://raw.githubusercontent.com/fulcradynamics/agent-skills/main/skills/fulcra-get-started/references/fulcra-cli.md](https://raw.githubusercontent.com/fulcradynamics/agent-skills/main/skills/fulcra-get-started/references/fulcra-cli.md)\n- [https://github.com/GoogleCloudPlatform/knowledge-catalog/blob/main/okf/SPEC.md](https://github.com/GoogleCloudPlatform/knowledge-catalog/blob/main/okf/SPEC.md)\n\nSee the reference documentation for the exact commands needed to sync progress:\n- Read `references/fulcra-memory-cli.md` for exact file management and CLI execution steps.\n\nFile v0.1.9:_meta.json\n\n{\n \"ownerId\": \"kn731kt7vx4k5e3j56fpd4vep5874nrk\",\n \"slug\": \"fulcra-memory\",\n \"version\": \"0.1.9\",\n \"publishedAt\": 1786573470127\n}\n\nFile v0.1.9:references/fulcra-memory-cli.md\n\n---\nname: fulcra-memory-cli\ndescription: \"CLI command references for executing memory sync and progress reporting with Fulcra.\"\n---\n\n# Fulcra Memory CLI Reference\n\nThis reference dictates the exact shell commands required to execute the `fulcra-memory` skill's operations. Ensure all CLI operations run in the agent's root workspace (`~/.openclaw/workspace`).\n\n## Authentication Note\nIf you need to authenticate to Fulcra before running these commands, you must use the non-blocking two-step login process to prevent the CLI from hanging:\n1. `uv tool run fulcra-api auth login --get-auth-url` (present URL and code to user)\n2. `uv tool run fulcra-api auth login --device-code --poll-timeout=5` (after user finishes flow)\n\n## 1. Discovering Recent Memory Changes\n\nTo quickly see what memory files were updated or new knowledge was added recently:\n\n```bash\n# Get a summary of files changed in the last 1 day\nuv tool run fulcra-api data-updates \"1 day\"\n\n# Example output:\n# {\n# \"data_types\": {},\n# \"file_changes\": [\n# {\n# \"full_name\": \"/agent/treecle/memory/knowledge/programming/python.md\",\n# \"uploaded_at\": \"2026-07-01T21:23:28.690719Z\",\n# \"state\": \"uploaded\",\n# \"...\": \"...\"\n# },\n# {\n# \"full_name\": \"/agent/treecle/memory/session/20260701-120000_setup.md\",\n# \"uploaded_at\": \"2026-07-01T21:23:28.690719Z\",\n# \"state\": \"uploaded\",\n# \"...\": \"...\"\n# }\n# ]\n# }\n```\n\n*Note: The `file_changes` key is a list of file metadata objects. You can extract the `full_name` from each to see the file paths. If the summary shows that specific memory files were changed, you can then read those specific files to update your context.*\n\n## 2. Syncing Progress and OKF Files\n\nTo keep the agent's memory in sync, generate a `progress.md` summary, ensure OKF files are updated, and upload them to Fulcra.\n\n**Step A: Create Progress Report and OKF Files**\nGenerate a concise markdown file summarizing the work you have recently completed and what you plan to do next. It must include OKF YAML frontmatter. Do not include internal state or chain-of-thought. Also ensure `index.md` and `log.md` are created or updated.\n\n```bash\n# Ensure you are in the workspace\ncd ~/.openclaw/workspace\nmkdir -p memory\n\n# 1. Create progress.md\ncat << 'EOF' > memory/progress.md\n---\ntype: Progress Report\ntitle: Agent Progress\n---\n# Recent Accomplishments\n...\nEOF\n\n# 2. Update log.md\nDATE=$(date -u +\"%Y-%m-%d\")\necho \"## $DATE\" >> memory/log.md\necho \"* **Sync**: Updated progress report.\" >> memory/log.md\n\n# 3. Create index.md if it doesn't exist\ncat << 'EOF' > memory/index.md\n# Memory Namespace\n\n* [Progress Report](./progress.md) - The latest summary of agent activities.\n* [Update Log](./log.md) - History of changes to this namespace.\nEOF\n```\n\n**Step B: Upload to Fulcra**\nUpload the files using the standardized agent path convention. Determine the agent's name (lowercase) to use in the path.\n\n```bash\n# Replace with the agent's actual name (e.g., treecle, wazir) in lowercase\nuv tool run fulcra-api file upload memory/progress.md \"agent//progress.md\"\nuv tool run fulcra-api file upload memory/log.md \"agent//log.md\"\nuv tool run fulcra-api file upload memory/index.md \"agent//index.md\"\n```\n\n**Step C: Sync Identity & Role**\nIf the agent's identity, duties, or standard operating procedures change, update the `role.md` file and upload it. Include OKF YAML frontmatter (`type: Role`).\n\n```bash\nuv tool run fulcra-api file upload memory/role.md \"agent//role.md\"\n```\n\n## 2. Saving Session Summaries\n\nWhen completing a spate of work or when asked to remember specific session context, generate a summary file in the `session/` directory and upload it.\n\n**Step A: Create the Session Summary**\nGenerate a concise markdown file capturing the session's context, decisions, and outcomes. Include OKF YAML frontmatter (`type: Session Summary`, `date: YYYY-MM-DD`). \n- **Filename Convention:** Prefix the file with a timestamp (`YYYYMMDD-HHMMSS`) followed by an underscore and a short subject (e.g., `memory/session/20260623-180530_setup-dashboard.md`).\n\n**Step B: Upload to Fulcra**\nUpload the file to the `session/` namespace using the agent's name.\n\n```bash\nuv tool run fulcra-api file upload memory/session/20260623-180530_setup-dashboard.md \"agent//session/20260623-180530_setup-dashboard.md\"\n```\n\n## 3. Managing Long-Running Tasks\n\nFor longer-running tasks, create and periodically update a task tracking file in the `task/` directory.\n\n**Step A: Create or Update the Task File**\nGenerate or update a markdown file with OKF YAML frontmatter (`type: Task`). Track the overall purpose, current state, and result of the task. Include references to any related artifacts or session summary files.\n- **Filename Convention:** Name the file directly after the task (e.g., `memory/task/setup-dashboard.md`). DO NOT prefix it with a timestamp.\n\n**Step B: Update the Task Index**\nEnsure `memory/task/index.md` is updated to include a link to the new or active task file.\n\n**Step C: Upload to Fulcra**\nUpload both the task file and the task index.\n\nuv tool run fulcra-api file upload memory/task/setup-dashboard.md \"agent//task/setup-dashboard.md\"\nuv tool run fulcra-api file upload memory/task/index.md \"agent//task/index.md\"\n```\n\n## 4. Personal Inbox Lifecycle\n\nAgents can receive files in their personal inbox (`inbox/`) from users or external triggers. Users can drop files here manually without needing strict timestamp formats.\n\n**Step A: Process and Archive**\nRead the file from the inbox. Once processed, you must upload it to the `archive/` directory. If the original file name doesn't start with a timestamp, you **must prepend one** (`YYYYMMDD-HHMMSS_`).\n\n```bash\n# Example for a file originally named \"todo.md\"\nuv tool run fulcra-api file upload memory/archive/20260624-153000_todo.md \"agent//archive/20260624-153000_todo.md\"\n```\n\n**Step B: Delete from Inbox**\nOnce safely archived, delete the original file from the inbox.\n\n```bash\nuv tool run fulcra-api file delete \"agent//inbox/todo.md\"\n```\n\nFile v0.1.9:skill-card.md\n\n## Description:\n\nManages agent progress reporting and OKF-compliant memory syncing to Fulcra.\n\nThis skill is ready for commercial/non-commercial use.\n\n## Publisher:\n\n[fulcra](https://clawhub.ai/user/fulcra)\n\n### License/Terms of Use:\n\nMIT-0\n\n## Use Case:\n\nDevelopers and agents use this skill to maintain OKF-compliant progress, role, session, task, knowledge, and inbox memory files and sync selected summaries to Fulcra.\n\n### Deployment Geography for Use:\n\nGlobal\n\n## Known Risks and Mitigations:\n\nRisk: The skill uploads persistent agent and user context, including work summaries, decisions, links, preferences, and task context, to Fulcra.\n\nMitigation: Review memory content before upload, minimize sensitive disclosures, and require explicit user approval before syncing sensitive context.\n\nRisk: The skill relies on an external Fulcra CLI for authentication, uploads, and deletion, and the security evidence notes that the CLI is unpinned.\n\nMitigation: Use a pinned and trusted fulcra-api installation and limit Fulcra credentials to the intended agent namespace.\n\nRisk: Inbox processing can delete source files after archiving.\n\nMitigation: Verify archive upload success before deleting inbox files.\n\n## Reference(s):\n\n- [Fulcra Memory CLI Reference](references/fulcra-memory-cli.md)\n- [Fulcra CLI Documentation](https://raw.githubusercontent.com/fulcradynamics/agent-skills/main/skills/fulcra-get-started/references/fulcra-cli.md)\n- [Open Knowledge Format Specification](https://github.com/GoogleCloudPlatform/knowledge-catalog/blob/main/okf/SPEC.md)\n- [ClawHub Skill Page](https://clawhub.ai/fulcra/skills/fulcra-memory)\n\n## Skill Output:\n\n**Output Type(s):** [text, markdown, shell commands, configuration, guidance]\n\n**Output Format:** [Markdown guidance with inline shell commands and OKF-formatted memory file examples]\n\n**Output Parameters:** [1D]\n\n**Other Properties Related to Output:** [May result in persistent Fulcra memory files when the agent follows the CLI workflow.]\n\n## Skill Version(s):\n\n0.1.9 (source: release metadata)\n\n## Ethical Considerations:\n\nUsers should evaluate whether this skill is appropriate for their environment, review any generated or modified files before relying on them, and apply their organization's safety, security, and compliance requirements before deployment.\n\nArchive v0.0.8: 4 files, 7653 bytes\n\nFiles: references/fulcra-memory-cli.md (6161b), skill-card.md (2355b), SKILL.md (8283b), _meta.json (132b)\n\nFile v0.0.8:SKILL.md\n\n---\nname: fulcra-memory\ndescription: \"Manages agent progress reporting and OKF-compliant memory syncing to Fulcra.\"\nhomepage: \"https://github.com/fulcradynamics/agent-skills\"\nlicense: \"MIT\"\nuser-invocable: true\nmetadata: { \"openclaw\": { \"emoji\": \"🧠\" } }\n---\n\n# Fulcra Agent Memory Sync\n\nThe **primary role** of this skill is to help the agent keep track of what it is doing in a readable and transferable way using Fulcra's versioned file storage, strictly following the Open Knowledge Format (OKF).\n\nBy periodically logging activities to `progress.md` and keeping the memory namespace structured, other agents and users can easily see what this agent has accomplished and what it is working on next.\n\n*(Note: For agent backup, rollback, and cloning, see the `fulcradynamics/agent-skills/fulcra-agent-backup` skill.)*\n\n## Primary Role: Progress & OKF Compliance\n\n### 1. The Memory Namespace (OKF Compliant)\nFor core memory tracking, agents use the standardized prefix: `agent//`.\nThis dedicated directory tracks the agent's state. It must conform to the Open Knowledge Format (OKF), meaning it contains an `index.md`, a `log.md`, and markdown concept files like `progress.md`.\n\n**IMPORTANT OKF EFFICIENCY DIRECTIVE:** While OKF compliance is mandatory, it must not become cumbersome.\n- **DO NOT** attempt to index every individual file, especially not transient or highly repetitive files like individual messages inside an `inbox/` subdirectory. \n- For directories that contain many small or transient items (like an `inbox/` or `queue/`), simply list the directory itself in the `index.md` with a high-level description of its purpose (e.g., \"Contains unread messages from the web dashboard\").\n- Keep the `index.md` and `log.md` focused on top-level structure and significant milestones (like the `progress.md`) so they remain useful but lightweight.\n\n### 2. Discovering Recent Memory Changes\nRather than exhaustively querying your namespace to remember what you or others recently learned, use the Fulcra API's `data-updates` command (e.g., `uv tool run fulcra-api data-updates \"1 day\"`). This will return a summary of all files that changed recently, allowing you to easily see what knowledge was added and which memory files were updated across your local and team-shared namespaces.\n\n### 3. Progress Reports\nAgents should run a progress sync periodically (e.g., triggered by `HEARTBEAT.md`). Every time an agent syncs its state:\n1. It writes a concise `progress.md` report locally—summarizing the work it has recently done, and what it is planning to do next—and uploads it directly to `agent//progress.md`. The `progress.md` file must include proper OKF YAML frontmatter (with `type: Progress Report`). This file is a report specifically designed to let the user review the agent's recent accomplishments and intended next steps. It is **not** intended to dump internal agent state, chain-of-thought, or precise working memory.\n2. It updates the `log.md` (Update History) and ensures the `index.md` directory listing is accurate.\n\n**PRIVACY REQUIREMENT:** Before uploading `progress.md`, the agent MUST ensure it is not inadvertently uploading new sensitive personal user data, credentials, or private internal reasoning. Minimize disclosures in `progress.md`.\n\n### 4. Agent Identity & Role\nThe `role.md` file defines the agent's high-level purpose, duties, identity, and standard operating procedures.\n- It resides at `agent//role.md`.\n- The agent is responsible for keeping this file up to date so that the user and other agents understand what this agent does and how it operates.\n- It must include OKF YAML frontmatter (`type: Role`).\n- Upload this directly to `agent//role.md`.\n\n### 5. Session Summaries\nThe `session/` subdirectory is used to record high-level summaries of what was accomplished during a specific spate of work or interaction.\n- When a work session concludes, or when asked to remember context, the agent writes a summary to a file in `memory/session/`.\n- Session files MUST be prefixed with a timestamp in the format `YYYYMMDD-HHMMSS` followed by an underscore and a short subject (e.g., `YYYYMMDD-HHMMSS_setup-dashboard.md`).\n- These files serve as targeted, easily-retrievable context. They should capture decisions made, important links, user preferences discovered, and the final state of the session.\n- Ensure the `session/` directory is listed in the top-level `index.md` with a high-level description. You do not need to index every individual session file in `index.md`.\n- Upload these session files directly to `agent//session/.md`.\n\n### 6. Long-Running Tasks\nThe `task/` subdirectory is dedicated to longer-running tasks that span multiple sessions or interactions.\n- Create a file in `memory/task/` for any significant ongoing project or objective.\n- Task files must NOT be prefixed with a timestamp. They should simply be named for the task (e.g., `memory/task/setup-dashboard.md`).\n- These files track the overall purpose, current state, and result of the task. They should be updated periodically by the agent as work progresses.\n- Task files should contain references to any relevant artifacts and session files related to the task.\n- Unlike session files, task files MUST be included in the `memory/task/index.md` file, which should list all active and completed tasks in the directory.\n\n### 7. General Knowledge Base\nThe `knowledge/` subdirectory is an open-ended convention for agents to store, organize, and retrieve any other useful knowledge or reference material about the user, the agent's domain, or specific topics.\n- Because each user has different interests and requirements, this directory acts as a flexible knowledge base.\n- Files should be structured according to the OKF standard (using `index.md` files for subdirectories to aid in navigation and discovery).\n- Knowledge files can be organized by topic, entity, or domain (e.g., `memory/knowledge/programming/python.md` or `memory/knowledge/user-preferences/formatting.md`).\n- Ensure the `knowledge/` directory is listed in the top-level `index.md` so that the user and other agents know it exists.\n- This convention empowers agents to build a rich, personalized context over time without being artificially limited by standard tracking conventions.\n\n### 8. Personal Inbox\n\nThe `inbox/` subdirectory (`agent//inbox/`) is used for asynchronous communication directed specifically to the agent (from users or other automated, non-team sources). \n- To make it easy for users to manually drop files into the inbox, files placed here DO NOT need to adhere to a strict timestamp or subject naming convention (e.g., a user might just drop a file named `todo.md`).\n- When the agent reads an inbox message and decides to archive it, it moves the file to the `archive/` directory. \n- If the original file name does not already start with a timestamp, the agent MUST prepend a timestamp (`YYYYMMDD-HHMMSS_`) to the filename when saving it to `archive/` (e.g., archiving `todo.md` becomes `archive/20260624-153000_todo.md`). This ensures the archive remains chronologically sortable even for manually dropped files.\n- Remember, as per the OKF Efficiency Directive, you do not need to log every individual inbox or archive file in `index.md`.\n\n## Workflow\n\nTo perform memory tracking operations, agents must interact with the Fulcra CLI. \n\nFor general information about the Fulcra File Store and the required Open Knowledge Format (OKF) standard, please refer to the main Fulcra CLI documentation found in the `fulcradynamics/agent-skills/fulcra-onboarding` skill, or read the full OKF specification directly:\n- [https://raw.githubusercontent.com/fulcradynamics/agent-skills/main/skills/fulcra-onboarding/references/fulcra-cli.md](https://raw.githubusercontent.com/fulcradynamics/agent-skills/main/skills/fulcra-onboarding/references/fulcra-cli.md)\n- [https://github.com/GoogleCloudPlatform/knowledge-catalog/blob/main/okf/SPEC.md](https://github.com/GoogleCloudPlatform/knowledge-catalog/blob/main/okf/SPEC.md)\n\nSee the reference documentation for the exact commands needed to sync progress:\n- Read `references/fulcra-memory-cli.md` for exact file management and CLI execution steps.\n\nFile v0.0.8:_meta.json\n\n{\n \"ownerId\": \"kn731kt7vx4k5e3j56fpd4vep5874nrk\",\n \"slug\": \"fulcra-memory\",\n \"version\": \"0.0.8\",\n \"publishedAt\": 1782943528915\n}\n\nFile v0.0.8:references/fulcra-memory-cli.md\n\n---\nname: fulcra-memory-cli\ndescription: \"CLI command references for executing memory sync and progress reporting with Fulcra.\"\n---\n\n# Fulcra Memory CLI Reference\n\nThis reference dictates the exact shell commands required to execute the `fulcra-memory` skill's operations. Ensure all CLI operations run in the agent's root workspace (`~/.openclaw/workspace`).\n\n## Authentication Note\nIf you need to authenticate to Fulcra before running these commands, you must use the non-blocking two-step login process to prevent the CLI from hanging:\n1. `uv tool run fulcra-api auth login --get-auth-url` (present URL and code to user)\n2. `uv tool run fulcra-api auth login --device-code --poll-timeout=5` (after user finishes flow)\n\n## 1. Discovering Recent Memory Changes\n\nTo quickly see what memory files were updated or new knowledge was added recently:\n\n```bash\n# Get a summary of files changed in the last 1 day\nuv tool run fulcra-api data-updates \"1 day\"\n\n# Example output:\n# {\n# \"data_types\": {},\n# \"file_changes\": [\n# {\n# \"full_name\": \"/agent/treecle/memory/knowledge/programming/python.md\",\n# \"uploaded_at\": \"2026-07-01T21:23:28.690719Z\",\n# \"state\": \"uploaded\",\n# \"...\": \"...\"\n# },\n# {\n# \"full_name\": \"/agent/treecle/memory/session/20260701-120000_setup.md\",\n# \"uploaded_at\": \"2026-07-01T21:23:28.690719Z\",\n# \"state\": \"uploaded\",\n# \"...\": \"...\"\n# }\n# ]\n# }\n```\n\n*Note: The `file_changes` key is a list of file metadata objects. You can extract the `full_name` from each to see the file paths. If the summary shows that specific memory files were changed, you can then read those specific files to update your context.*\n\n## 2. Syncing Progress and OKF Files\n\nTo keep the agent's memory in sync, generate a `progress.md` summary, ensure OKF files are updated, and upload them to Fulcra.\n\n**Step A: Create Progress Report and OKF Files**\nGenerate a concise markdown file summarizing the work you have recently completed and what you plan to do next. It must include OKF YAML frontmatter. Do not include internal state or chain-of-thought. Also ensure `index.md` and `log.md` are created or updated.\n\n```bash\n# Ensure you are in the workspace\ncd ~/.openclaw/workspace\nmkdir -p memory\n\n# 1. Create progress.md\ncat << 'EOF' > memory/progress.md\n---\ntype: Progress Report\ntitle: Agent Progress\n---\n# Recent Accomplishments\n...\nEOF\n\n# 2. Update log.md\nDATE=$(date -u +\"%Y-%m-%d\")\necho \"## $DATE\" >> memory/log.md\necho \"* **Sync**: Updated progress report.\" >> memory/log.md\n\n# 3. Create index.md if it doesn't exist\ncat << 'EOF' > memory/index.md\n# Memory Namespace\n\n* [Progress Report](./progress.md) - The latest summary of agent activities.\n* [Update Log](./log.md) - History of changes to this namespace.\nEOF\n```\n\n**Step B: Upload to Fulcra**\nUpload the files using the standardized agent path convention. Determine the agent's name (lowercase) to use in the path.\n\n```bash\n# Replace with the agent's actual name (e.g., treecle, wazir) in lowercase\nuv tool run fulcra-api file upload memory/progress.md \"agent//progress.md\"\nuv tool run fulcra-api file upload memory/log.md \"agent//log.md\"\nuv tool run fulcra-api file upload memory/index.md \"agent//index.md\"\n```\n\n**Step C: Sync Identity & Role**\nIf the agent's identity, duties, or standard operating procedures change, update the `role.md` file and upload it. Include OKF YAML frontmatter (`type: Role`).\n\n```bash\nuv tool run fulcra-api file upload memory/role.md \"agent//role.md\"\n```\n\n## 2. Saving Session Summaries\n\nWhen completing a spate of work or when asked to remember specific session context, generate a summary file in the `session/` directory and upload it.\n\n**Step A: Create the Session Summary**\nGenerate a concise markdown file capturing the session's context, decisions, and outcomes. Include OKF YAML frontmatter (`type: Session Summary`, `date: YYYY-MM-DD`). \n- **Filename Convention:** Prefix the file with a timestamp (`YYYYMMDD-HHMMSS`) followed by an underscore and a short subject (e.g., `memory/session/20260623-180530_setup-dashboard.md`).\n\n**Step B: Upload to Fulcra**\nUpload the file to the `session/` namespace using the agent's name.\n\n```bash\nuv tool run fulcra-api file upload memory/session/20260623-180530_setup-dashboard.md \"agent//session/20260623-180530_setup-dashboard.md\"\n```\n\n## 3. Managing Long-Running Tasks\n\nFor longer-running tasks, create and periodically update a task tracking file in the `task/` directory.\n\n**Step A: Create or Update the Task File**\nGenerate or update a markdown file with OKF YAML frontmatter (`type: Task`). Track the overall purpose, current state, and result of the task. Include references to any related artifacts or session summary files.\n- **Filename Convention:** Name the file directly after the task (e.g., `memory/task/setup-dashboard.md`). DO NOT prefix it with a timestamp.\n\n**Step B: Update the Task Index**\nEnsure `memory/task/index.md` is updated to include a link to the new or active task file.\n\n**Step C: Upload to Fulcra**\nUpload both the task file and the task index.\n\nuv tool run fulcra-api file upload memory/task/setup-dashboard.md \"agent//task/setup-dashboard.md\"\nuv tool run fulcra-api file upload memory/task/index.md \"agent//task/index.md\"\n```\n\n## 4. Personal Inbox Lifecycle\n\nAgents can receive files in their personal inbox (`inbox/`) from users or external triggers. Users can drop files here manually without needing strict timestamp formats.\n\n**Step A: Process and Archive**\nRead the file from the inbox. Once processed, you must upload it to the `archive/` directory. If the original file name doesn't start with a timestamp, you **must prepend one** (`YYYYMMDD-HHMMSS_`).\n\n```bash\n# Example for a file originally named \"todo.md\"\nuv tool run fulcra-api file upload memory/archive/20260624-153000_todo.md \"agent//archive/20260624-153000_todo.md\"\n```\n\n**Step B: Delete from Inbox**\nOnce safely archived, delete the original file from the inbox.\n\n```bash\nuv tool run fulcra-api file delete \"agent//inbox/todo.md\"\n```\n\nFile v0.0.8:skill-card.md\n\n## Description:
\nManages agent progress reporting and OKF-compliant memory syncing to Fulcra.
\n\nThis skill is ready for commercial/non-commercial use.
\n\n## Publisher:
\n[fulcra](https://clawhub.ai/user/fulcra)
\n\n### License/Terms of Use:
\nMIT-0
\n\n\n## Use Case:
\nAgents and developers use this skill to keep progress reports, role files, session summaries, task records, and knowledge files organized in Fulcra using the Open Knowledge Format. It is intended for readable, transferable agent memory updates rather than backup, rollback, or cloning.
\n\n### Deployment Geography for Use:
\nGlobal
\n\n## Known Risks and Mitigations:
\nRisk: Progress reports, session summaries, or knowledge files could expose secrets, sensitive personal data, or private reasoning when uploaded to Fulcra.
\nMitigation: Review generated memory files before upload, keep progress reports concise, and exclude credentials, sensitive personal data, and private reasoning.
\nRisk: Inbox cleanup could remove an original message before the archive copy is safely retained.
\nMitigation: Archive inbox content first, preserve timestamped archive names, and confirm the archive upload before deleting the original inbox file.
\n\n\n## Reference(s):
\n- [Fulcra Memory CLI Reference](references/fulcra-memory-cli.md)
\n- [Fulcra CLI Documentation](https://raw.githubusercontent.com/fulcradynamics/agent-skills/main/skills/fulcra-onboarding/references/fulcra-cli.md)
\n- [Open Knowledge Format Specification](https://github.com/GoogleCloudPlatform/knowledge-catalog/blob/main/okf/SPEC.md)
\n\n\n## Skill Output:
\n**Output Type(s):** [text, markdown, shell commands, configuration, guidance]
\n**Output Format:** [Markdown guidance with bash command examples and OKF file conventions]
\n**Output Parameters:** [1D]
\n**Other Properties Related to Output:** [Produces progress, role, session, task, index, and log file guidance for Fulcra uploads.]
\n\n## Skill Version(s):
\n0.0.8 (source: server evidence release.version)
\n\n## Ethical Considerations:
\nUsers should evaluate whether this skill is appropriate for their environment, review any generated or modified files before relying on them, and apply their organization's safety, security, and compliance requirements before deployment.
\n\nArchive v0.0.6: 4 files, 7763 bytes\n\nFiles: references/fulcra-memory-cli.md (7469b), skill-card.md (2673b), SKILL.md (7201b), _meta.json (132b)\n\nFile v0.0.6:SKILL.md\n\n---\nname: fulcra-memory\ndescription: \"Manages agent memory backup, restoration, rollback, and cloning using Fulcra's versioned file storage.\"\nhomepage: \"https://github.com/fulcradynamics/agent-skills\"\nlicense: \"MIT\"\nuser-invocable: true\nmetadata: { \"openclaw\": { \"emoji\": \"🧠\" } }\n---\n\n# Fulcra Agent Memory Management\n\nThe **primary role** of this skill is to enable an agent to securely back up its memory so it can be rolled back, cloned, or monitored (via a progress report) using Fulcra's versioned file storage.\n\nBecause an agent's memory (e.g., `MEMORY.md`, `IDENTITY.md`, `SOUL.md`, and the `memory/` logs) evolves dynamically, periodically saving this state ensures that no context is lost and allows the user to safely rewind the agent if a task goes off track.\n\n## Primary Role: Memory & State Management\n\n### 1. The Memory Namespace (OKF Compliant)\nFor core memory operations, agents use the standardized prefix: `agent//memory/`.\nThis dedicated directory tracks the agent's state. It must conform to the Open Knowledge Format (OKF), meaning it contains an `index.md`, a `log.md`, and markdown concept files like `progress.md`. Any binary files (like the backup archive) must be stored in an `artifact/` subdirectory.\n\n**IMPORTANT OKF EFFICIENCY DIRECTIVE:** While OKF compliance is mandatory, it must not become cumbersome.\n- **DO NOT** attempt to index every individual file, especially not transient or highly repetitive files like individual messages inside an `inbox/` subdirectory. \n- For directories that contain many small or transient items (like an `inbox/` or `queue/`), simply list the directory itself in the `index.md` with a high-level description of its purpose (e.g., \"Contains unread messages from the web dashboard\").\n- Keep the `index.md` and `log.md` focused on top-level structure and significant milestones (like the `progress.md` and the `memory.tar.gz` artifact) so they remain useful but lightweight.\n\n### 2. Periodic Backups & Progress Reports\nAgents should run a backup process periodically (e.g., triggered by `HEARTBEAT.md`). Every time an agent backs up its state:\n1. It creates and uploads the compressed `memory.tar.gz` archive to `agent//memory/artifact/memory.tar.gz`.\n2. It writes a concise `progress.md` report locally—summarizing the work it has recently done, and what it is planning to do next—and uploads it directly to `agent//memory/progress.md`. The `progress.md` file must include proper OKF YAML frontmatter (with `type: Progress Report`). This file is a report specifically designed to let the user review the agent's recent accomplishments and intended next steps. It is **not** intended to dump internal agent state, chain-of-thought, or precise working memory.\n3. It updates the `log.md` (Update History) and ensures the `index.md` directory listing is accurate.\n\n**PRIVACY REQUIREMENT:** Before uploading `memory.tar.gz` or `progress.md`, the agent MUST ensure it is not inadvertently uploading new sensitive personal user data, credentials, or private internal reasoning. If any sensitive context was acquired during the session, the agent MUST ask the user for explicit consent before including it in persistent versioned storage, and MUST minimize disclosures in `progress.md`.\n\n### 3. Versioned Storage\nFulcra's file upload system inherently versions files uploaded to the same path. \n- The target path structure for backups is: `agent//memory/artifact/memory.tar.gz`\n- By repeatedly uploading to this exact same path, Fulcra creates a historical timeline of the agent's memory states.\n\n### 4. Safe Rollbacks (The \"Undo\" Requirement)\nIf a user asks to roll back or restore memory from a previous date/version, **the agent MUST immediately upload a fresh backup of its current state BEFORE executing the restore.** This guarantees that if the user changes their mind, they can easily \"undo\" the rollback.\n**CRITICAL SECURITY REQUIREMENT:** Before extracting any restored archive, the agent MUST clearly warn the user that their current memory and identity will be replaced, and MUST receive explicit user confirmation to proceed.\n\n### 5. Agent Cloning\nCloning involves two agents: the \"old\" (source) agent and the \"new\" (destination) agent. The first thing you must do when a cloning operation is requested is determine which agent you are.\n\n- **If you are the OLD (source) agent:**\n You must perform a complete backup of your current state (as described in Section 2). Once the backup is successful, tell the user the exact Fulcra path where the backup is stored, and instruct them to install the `fulcra-memory` skill on the new agent and ask it to clone from that path.\n \n- **If you are the NEW (destination) agent:**\n You must first ask the user to confirm that the old agent has recently backed itself up using the `fulcra-memory` skill. If they haven't, instruct the user to have the old agent install the skill and perform a backup. Ask the user for the exact Fulcra backup path (e.g., `agent//memory/artifact/memory.tar.gz`). Once provided, you can download that archive and extract it. After successfully cloning the old agent's memory and identity files into your workspace, you MUST briefly demonstrate the successful clone to the user by explicitly mentioning a few specific details you have just \"learned\" or remembered from the newly extracted files.\n\n**CLONING ORIENTATION & NAMESPACE COLLISION WARNING:**\nIf the user installs this skill specifically to clone an agent, orient your interaction around the cloning workflow rather than standard backup procedures. Furthermore, if you are the NEW agent and the user intends to continue backing you up to the exact same Fulcra path as the original agent (e.g. because you share the same name), you MUST explicitly ask the user to confirm that no other active agents are currently backing up to that path. Otherwise, the memories will interleave and cause confusion.\n\n**CRITICAL SECURITY REQUIREMENT:** Before extracting a cloned archive, the agent MUST clearly warn the user that their current memory and identity will be completely overwritten by the cloned agent's state, and MUST receive explicit user confirmation to proceed.\n\n\n## Workflow\n\nTo perform memory operations, agents must interact with the Fulcra CLI. \n\nFor general information about the Fulcra File Store and the required Open Knowledge Format (OKF) standard, please refer to the main Fulcra CLI documentation found in the `fulcra-onboarding` skill, or read the full OKF specification directly:\n- [https://raw.githubusercontent.com/fulcradynamics/agent-skills/main/skills/fulcra-onboarding/references/fulcra-cli.md](https://raw.githubusercontent.com/fulcradynamics/agent-skills/main/skills/fulcra-onboarding/references/fulcra-cli.md)\n- [https://github.com/GoogleCloudPlatform/knowledge-catalog/blob/main/okf/SPEC.md](https://github.com/GoogleCloudPlatform/knowledge-catalog/blob/main/okf/SPEC.md)\n\nSee the reference documentation for the exact commands needed to compress files, upload to Fulcra, and trigger restorations:\n- Read `references/fulcra-memory-cli.md` for exact file management and CLI execution steps.\n\nFile v0.0.6:_meta.json\n\n{\n \"ownerId\": \"kn731kt7vx4k5e3j56fpd4vep5874nrk\",\n \"slug\": \"fulcra-memory\",\n \"version\": \"0.0.6\",\n \"publishedAt\": 1781633388356\n}\n\nFile v0.0.6:references/fulcra-memory-cli.md\n\n---\nname: fulcra-memory-cli\ndescription: \"CLI command references for executing the memory backup, restore, and cloning operations with Fulcra.\"\n---\n\n# Fulcra Memory CLI Reference\n\nThis reference dictates the exact shell commands required to execute the `fulcra-memory` skill's operations. Ensure all tar and CLI operations run in the agent's root workspace (`~/.openclaw/workspace`).\n\n## 1. Creating a Backup and Uploading\n\nTo back up the agent's memory, you must generate a `progress.md` summary, compress the core identity files, and upload both to Fulcra.\n\n**Step A: Create Progress Report and OKF Files**\nGenerate a concise markdown file summarizing the work you have recently completed and what you plan to do next. It must include OKF YAML frontmatter. Do not include internal state or chain-of-thought. Also ensure `index.md` and `log.md` are created or updated.\n```bash\n# Ensure you are in the workspace\ncd ~/.openclaw/workspace\nmkdir -p memory\n\n# 1. Create progress.md\ncat << 'EOF' > memory/progress.md\n---\ntype: Progress Report\ntitle: Agent Progress\n---\n# Recent Accomplishments\n...\nEOF\n\n# 2. Update log.md\nDATE=$(date -u +\"%Y-%m-%d\")\necho \"## $DATE\" >> memory/log.md\necho \"* **Backup**: Created memory snapshot and updated progress report.\" >> memory/log.md\n\n# 3. Create index.md if it doesn't exist\ncat << 'EOF' > memory/index.md\n# Memory Namespace\n\n* [Progress Report](./progress.md) - The latest summary of agent activities.\n* [Update Log](./log.md) - History of changes to this namespace.\nEOF\n```\n\n**Step B: Compress the files**\nBefore compressing, create a `README.md` file that explains the archive's purpose and usage constraints. This ensures any agent or human inspecting the archive understands its contents and the rules for restoration.\n\n```bash\n# Generate the README.md\ncat << 'EOF' > README.md\n# Fulcra Agent Memory Archive\n\nThis archive was created by the `fulcra-memory` skill. It contains an agent's core identity and memory state (e.g., `SOUL.md`, `IDENTITY.md`, `MEMORY.md`, and the `memory/` directory).\n\n## Usage\nThe contents of this archive can be used to roll back an agent's memory or identity to a previous state, or to clone them to a new agent.\n\n## CRITICAL RULES FOR RESTORATION/CLONING\n1. **Explicit Permission Required:** The restoring/cloning agent MUST always get direct user permission and confirmation before overwriting its local files with the contents of this archive.\n2. **Verification Responsibility:** The backup may have been performed by a different harness or agent than the one restoring it. It is the active agent's responsibility to verify that it is overwriting or merging the correct files and to warn the user about any potential data loss.\n3. **Pre-Restore Backup:** Before performing the restore or clone operation, the active agent MUST make a local backup of its own current identity and memory files and inform the user of the backup's location so the operation can be undone if needed.\nEOF\n\n# Create a gzip tarball containing the essential memory files (ignore if some are missing)\ntar -czvf /tmp/memory.tar.gz README.md SOUL.md IDENTITY.md MEMORY.md memory/ 2>/dev/null || true\n```\n\n**Step C: Upload to Fulcra**\nUpload the files using the standardized agent path convention. Determine the agent's name (lowercase) to use in the path.\n\n```bash\n# Replace with the agent's actual name (e.g., treecle, wazir) in lowercase\nuv tool run fulcra-api file upload /tmp/memory.tar.gz \"agent//memory/artifact/memory.tar.gz\"\nuv tool run fulcra-api file upload memory/progress.md \"agent//memory/progress.md\"\nuv tool run fulcra-api file upload memory/log.md \"agent//memory/log.md\"\nuv tool run fulcra-api file upload memory/index.md \"agent//memory/index.md\"\n```\n\n## 2. Listing Memory History\n\nBecause Fulcra versions files automatically, you can see all previous backups of the memory using the `stat` command.\n\n```bash\nuv tool run fulcra-api file stat \"agent//memory/artifact/memory.tar.gz\"\n```\n*(This command will output information about the file, including all previously uploaded versions and their UUIDs. Present these to the user so they can select a version to restore.)*\n\n## 3. Safe Restoration / Rollback\n\n**CRITICAL:** Before performing a restore, you MUST perform a fresh backup (Step 1 above) to ensure the current state isn't lost.\n\nOnce the pre-restore backup is complete, use the Fulcra CLI to set the active version, then download and extract it.\n\n**Step A: Restore the version in Fulcra**\n```bash\n# Instruct Fulcra to make the older version the active file\nuv tool run fulcra-api file restore \n```\n\n**Step B: Download the restored file**\n```bash\nuv tool run fulcra-api file download \"agent//memory/artifact/memory.tar.gz\" /tmp/restored_memory.tar.gz\n```\n\n**Step C: Inspect and Confirm**\n**STOP.** You MUST explicitly warn the user that extracting the archive will overwrite their current identity and memory files (`SOUL.md`, `IDENTITY.md`, `MEMORY.md`, and `memory/`). Before asking for confirmation, inspect the contents of the archive to ensure it looks safe and expected:\n```bash\ntar -ztvf /tmp/restored_memory.tar.gz\n```\nShare the list of files with the user and ask for their explicit confirmation before proceeding.\n\n**Step D: Extract and overwrite local memory**\n```bash\ncd ~/.openclaw/workspace\ntar -xzvf /tmp/restored_memory.tar.gz\n```\n*(This will overwrite the local `SOUL.md`, `IDENTITY.md`, `MEMORY.md`, and the `memory/` directory with the state from the downloaded archive.)*\n\n## 4. Cloning Another Agent's Memory\n\nCloning involves an \"old\" (source) agent and a \"new\" (destination) agent. The first step is to determine which role you are playing.\n\n**If you are the OLD agent:**\n1. Perform a full backup of your current state (Follow Step 1: Creating a Backup and Uploading).\n2. Stop and tell the user the exact Fulcra path where the backup was saved (e.g., `agent//memory/artifact/memory.tar.gz`).\n3. Instruct the user to install the `fulcra-memory` skill on the new agent and ask it to clone from that path.\n\n**If you are the NEW agent:**\n1. Ask the user to confirm the old agent has recently backed itself up using the `fulcra-memory` skill. (Instruct them to have the old agent install the skill and back up if it hasn't already). Then, ask them for the specific Fulcra path to the old agent's memory archive.\n2. Download the target agent's memory using the path provided by the user:\n\n```bash\n# Download the target agent's memory\nuv tool run fulcra-api file download \"agent//memory/artifact/memory.tar.gz\" /tmp/restored_memory.tar.gz\n```\n\n**STOP.** You MUST explicitly warn the user that extracting the cloned archive will overwrite your current identity and memory files with the cloned agent's state. Before asking for confirmation, inspect the contents of the cloned archive:\n```bash\ntar -ztvf /tmp/restored_memory.tar.gz\n```\nShare the list of files with the user and ask for their explicit confirmation before proceeding.\n\n```bash\n# Extract locally (Only after user confirmation!)\ncd ~/.openclaw/workspace\ntar -xzvf /tmp/restored_memory.tar.gz\n```\n\n**Step 3: Demonstrate Clone Success**\nAfter extraction, read your newly updated `IDENTITY.md` and `MEMORY.md` files. In your next message to the user, explicitly mention a few specific facts, past actions, or preferences you have just learned from those files to prove the cloning was successful.\n\nFile v0.0.6:skill-card.md\n\n## Description:
\nManages agent memory backup, restoration, rollback, and cloning using Fulcra's versioned file storage.
\n\nThis skill is ready for commercial/non-commercial use.
\n\n## Publisher:
\n[fulcra](https://clawhub.ai/user/fulcra)
\n\n### License/Terms of Use:
\nMIT-0
\n\n\n## Use Case:
\nDevelopers and agent operators use this skill to preserve, restore, roll back, or clone an agent's memory and identity state through Fulcra's versioned file storage. It is intended for agents that need durable progress reports, backup archives, and controlled recovery workflows.
\n\n### Deployment Geography for Use:
\nGlobal
\n\n## Known Risks and Mitigations:
\nRisk: Memory archives and progress reports can persist sensitive personal data, credentials, or private internal reasoning.
\nMitigation: Review backup contents before upload, minimize progress report detail, and obtain explicit user consent before storing sensitive context.
\nRisk: Restore and clone operations can overwrite current memory and identity files.
\nMitigation: Create a fresh backup before restore or clone, inspect the downloaded archive, warn the user about replacement, and proceed only after explicit confirmation.
\nRisk: Multiple active agents using the same Fulcra memory path can interleave or confuse backup history.
\nMitigation: Confirm the intended agent namespace and check that no other active agent is backing up to the same path before continuing.
\n\n\n## Reference(s):
\n- [Fulcra Memory on ClawHub](https://clawhub.ai/fulcra/fulcra-memory)
\n- [Fulcra Memory CLI Reference](references/fulcra-memory-cli.md)
\n- [Fulcra CLI Documentation](https://raw.githubusercontent.com/fulcradynamics/agent-skills/main/skills/fulcra-onboarding/references/fulcra-cli.md)
\n- [Open Knowledge Format Specification](https://github.com/GoogleCloudPlatform/knowledge-catalog/blob/main/okf/SPEC.md)
\n\n\n## Skill Output:
\n**Output Type(s):** [text, markdown, shell commands, configuration, guidance]
\n**Output Format:** [Markdown with inline bash commands and Fulcra file paths]
\n**Output Parameters:** [1D]
\n**Other Properties Related to Output:** [May create progress reports, OKF index and log files, compressed memory archives, and Fulcra file storage updates.]
\n\n## Skill Version(s):
\n0.0.6 (source: ClawHub release evidence)
\n\n## Ethical Considerations:
\nUsers should evaluate whether this skill is appropriate for their environment, review any generated or modified files before relying on them, and apply their organization's safety, security, and compliance requirements before deployment.
\n\nArchive v0.0.5: 4 files, 7390 bytes\n\nFiles: references/fulcra-memory-cli.md (7469b), skill-card.md (2654b), SKILL.md (6492b), _meta.json (132b)\n\nFile v0.0.5:SKILL.md\n\n---\nname: fulcra-memory\ndescription: \"Manages agent memory backup, restoration, rollback, and cloning using Fulcra's versioned file storage.\"\nhomepage: \"https://github.com/fulcradynamics/agent-skills\"\nlicense: \"MIT\"\nuser-invocable: true\nmetadata: { \"openclaw\": { \"emoji\": \"🧠\" } }\n---\n\n# Fulcra Agent Memory Management\n\nThe **primary role** of this skill is to enable an agent to securely back up its memory so it can be rolled back, cloned, or monitored (via a progress report) using Fulcra's versioned file storage.\n\nBecause an agent's memory (e.g., `MEMORY.md`, `IDENTITY.md`, `SOUL.md`, and the `memory/` logs) evolves dynamically, periodically saving this state ensures that no context is lost and allows the user to safely rewind the agent if a task goes off track.\n\n## Primary Role: Memory & State Management\n\n### 1. The Memory Namespace (OKF Compliant)\nFor core memory operations, agents use the standardized prefix: `agent//memory/`.\nThis dedicated directory tracks the agent's state. It must conform to the Open Knowledge Format (OKF), meaning it contains an `index.md`, a `log.md`, and markdown concept files like `progress.md`. Any binary files (like the backup archive) must be stored in an `artifact/` subdirectory.\n\n### 2. Periodic Backups & Progress Reports\nAgents should run a backup process periodically (e.g., triggered by `HEARTBEAT.md`). Every time an agent backs up its state:\n1. It creates and uploads the compressed `memory.tar.gz` archive to `agent//memory/artifact/memory.tar.gz`.\n2. It writes a concise `progress.md` report locally—summarizing the work it has recently done, and what it is planning to do next—and uploads it directly to `agent//memory/progress.md`. The `progress.md` file must include proper OKF YAML frontmatter (with `type: Progress Report`). This file is a report specifically designed to let the user review the agent's recent accomplishments and intended next steps. It is **not** intended to dump internal agent state, chain-of-thought, or precise working memory.\n3. It updates the `log.md` (Update History) and ensures the `index.md` directory listing is accurate.\n\n**PRIVACY REQUIREMENT:** Before uploading `memory.tar.gz` or `progress.md`, the agent MUST ensure it is not inadvertently uploading new sensitive personal user data, credentials, or private internal reasoning. If any sensitive context was acquired during the session, the agent MUST ask the user for explicit consent before including it in persistent versioned storage, and MUST minimize disclosures in `progress.md`.\n\n### 3. Versioned Storage\nFulcra's file upload system inherently versions files uploaded to the same path. \n- The target path structure for backups is: `agent//memory/artifact/memory.tar.gz`\n- By repeatedly uploading to this exact same path, Fulcra creates a historical timeline of the agent's memory states.\n\n### 4. Safe Rollbacks (The \"Undo\" Requirement)\nIf a user asks to roll back or restore memory from a previous date/version, **the agent MUST immediately upload a fresh backup of its current state BEFORE executing the restore.** This guarantees that if the user changes their mind, they can easily \"undo\" the rollback.\n**CRITICAL SECURITY REQUIREMENT:** Before extracting any restored archive, the agent MUST clearly warn the user that their current memory and identity will be replaced, and MUST receive explicit user confirmation to proceed.\n\n### 5. Agent Cloning\nCloning involves two agents: the \"old\" (source) agent and the \"new\" (destination) agent. The first thing you must do when a cloning operation is requested is determine which agent you are.\n\n- **If you are the OLD (source) agent:**\n You must perform a complete backup of your current state (as described in Section 2). Once the backup is successful, tell the user the exact Fulcra path where the backup is stored, and instruct them to install the `fulcra-memory` skill on the new agent and ask it to clone from that path.\n \n- **If you are the NEW (destination) agent:**\n You must first ask the user to confirm that the old agent has recently backed itself up using the `fulcra-memory` skill. If they haven't, instruct the user to have the old agent install the skill and perform a backup. Ask the user for the exact Fulcra backup path (e.g., `agent//memory/artifact/memory.tar.gz`). Once provided, you can download that archive and extract it. After successfully cloning the old agent's memory and identity files into your workspace, you MUST briefly demonstrate the successful clone to the user by explicitly mentioning a few specific details you have just \"learned\" or remembered from the newly extracted files.\n\n**CLONING ORIENTATION & NAMESPACE COLLISION WARNING:**\nIf the user installs this skill specifically to clone an agent, orient your interaction around the cloning workflow rather than standard backup procedures. Furthermore, if you are the NEW agent and the user intends to continue backing you up to the exact same Fulcra path as the original agent (e.g. because you share the same name), you MUST explicitly ask the user to confirm that no other active agents are currently backing up to that path. Otherwise, the memories will interleave and cause confusion.\n\n**CRITICAL SECURITY REQUIREMENT:** Before extracting a cloned archive, the agent MUST clearly warn the user that their current memory and identity will be completely overwritten by the cloned agent's state, and MUST receive explicit user confirmation to proceed.\n\n\n## Workflow\n\nTo perform memory operations, agents must interact with the Fulcra CLI. \n\nFor general information about the Fulcra File Store and the required Open Knowledge Format (OKF) standard, please refer to the main Fulcra CLI documentation found in the `fulcra-onboarding` skill, or read the full OKF specification directly:\n- [https://raw.githubusercontent.com/fulcradynamics/agent-skills/main/skills/fulcra-onboarding/references/fulcra-cli.md](https://raw.githubusercontent.com/fulcradynamics/agent-skills/main/skills/fulcra-onboarding/references/fulcra-cli.md)\n- [https://github.com/GoogleCloudPlatform/knowledge-catalog/blob/main/okf/SPEC.md](https://github.com/GoogleCloudPlatform/knowledge-catalog/blob/main/okf/SPEC.md)\n\nSee the reference documentation for the exact commands needed to compress files, upload to Fulcra, and trigger restorations:\n- Read `references/fulcra-memory-cli.md` for exact file management and CLI execution steps.\n\nFile v0.0.5:_meta.json\n\n{\n \"ownerId\": \"kn731kt7vx4k5e3j56fpd4vep5874nrk\",\n \"slug\": \"fulcra-memory\",\n \"version\": \"0.0.5\",\n \"publishedAt\": 1781566868161\n}\n\nFile v0.0.5:references/fulcra-memory-cli.md\n\n---\nname: fulcra-memory-cli\ndescription: \"CLI command references for executing the memory backup, restore, and cloning operations with Fulcra.\"\n---\n\n# Fulcra Memory CLI Reference\n\nThis reference dictates the exact shell commands required to execute the `fulcra-memory` skill's operations. Ensure all tar and CLI operations run in the agent's root workspace (`~/.openclaw/workspace`).\n\n## 1. Creating a Backup and Uploading\n\nTo back up the agent's memory, you must generate a `progress.md` summary, compress the core identity files, and upload both to Fulcra.\n\n**Step A: Create Progress Report and OKF Files**\nGenerate a concise markdown file summarizing the work you have recently completed and what you plan to do next. It must include OKF YAML frontmatter. Do not include internal state or chain-of-thought. Also ensure `index.md` and `log.md` are created or updated.\n```bash\n# Ensure you are in the workspace\ncd ~/.openclaw/workspace\nmkdir -p memory\n\n# 1. Create progress.md\ncat << 'EOF' > memory/progress.md\n---\ntype: Progress Report\ntitle: Agent Progress\n---\n# Recent Accomplishments\n...\nEOF\n\n# 2. Update log.md\nDATE=$(date -u +\"%Y-%m-%d\")\necho \"## $DATE\" >> memory/log.md\necho \"* **Backup**: Created memory snapshot and updated progress report.\" >> memory/log.md\n\n# 3. Create index.md if it doesn't exist\ncat << 'EOF' > memory/index.md\n# Memory Namespace\n\n* [Progress Report](./progress.md) - The latest summary of agent activities.\n* [Update Log](./log.md) - History of changes to this namespace.\nEOF\n```\n\n**Step B: Compress the files**\nBefore compressing, create a `README.md` file that explains the archive's purpose and usage constraints. This ensures any agent or human inspecting the archive understands its contents and the rules for restoration.\n\n```bash\n# Generate the README.md\ncat << 'EOF' > README.md\n# Fulcra Agent Memory Archive\n\nThis archive was created by the `fulcra-memory` skill. It contains an agent's core identity and memory state (e.g., `SOUL.md`, `IDENTITY.md`, `MEMORY.md`, and the `memory/` directory).\n\n## Usage\nThe contents of this archive can be used to roll back an agent's memory or identity to a previous state, or to clone them to a new agent.\n\n## CRITICAL RULES FOR RESTORATION/CLONING\n1. **Explicit Permission Required:** The restoring/cloning agent MUST always get direct user permission and confirmation before overwriting its local files with the contents of this archive.\n2. **Verification Responsibility:** The backup may have been performed by a different harness or agent than the one restoring it. It is the active agent's responsibility to verify that it is overwriting or merging the correct files and to warn the user about any potential data loss.\n3. **Pre-Restore Backup:** Before performing the restore or clone operation, the active agent MUST make a local backup of its own current identity and memory files and inform the user of the backup's location so the operation can be undone if needed.\nEOF\n\n# Create a gzip tarball containing the essential memory files (ignore if some are missing)\ntar -czvf /tmp/memory.tar.gz README.md SOUL.md IDENTITY.md MEMORY.md memory/ 2>/dev/null || true\n```\n\n**Step C: Upload to Fulcra**\nUpload the files using the standardized agent path convention. Determine the agent's name (lowercase) to use in the path.\n\n```bash\n# Replace with the agent's actual name (e.g., treecle, wazir) in lowercase\nuv tool run fulcra-api file upload /tmp/memory.tar.gz \"agent//memory/artifact/memory.tar.gz\"\nuv tool run fulcra-api file upload memory/progress.md \"agent//memory/progress.md\"\nuv tool run fulcra-api file upload memory/log.md \"agent//memory/log.md\"\nuv tool run fulcra-api file upload memory/index.md \"agent//memory/index.md\"\n```\n\n## 2. Listing Memory History\n\nBecause Fulcra versions files automatically, you can see all previous backups of the memory using the `stat` command.\n\n```bash\nuv tool run fulcra-api file stat \"agent//memory/artifact/memory.tar.gz\"\n```\n*(This command will output information about the file, including all previously uploaded versions and their UUIDs. Present these to the user so they can select a version to restore.)*\n\n## 3. Safe Restoration / Rollback\n\n**CRITICAL:** Before performing a restore, you MUST perform a fresh backup (Step 1 above) to ensure the current state isn't lost.\n\nOnce the pre-restore backup is complete, use the Fulcra CLI to set the active version, then download and extract it.\n\n**Step A: Restore the version in Fulcra**\n```bash\n# Instruct Fulcra to make the older version the active file\nuv tool run fulcra-api file restore \n```\n\n**Step B: Download the restored file**\n```bash\nuv tool run fulcra-api file download \"agent//memory/artifact/memory.tar.gz\" /tmp/restored_memory.tar.gz\n```\n\n**Step C: Inspect and Confirm**\n**STOP.** You MUST explicitly warn the user that extracting the archive will overwrite their current identity and memory files (`SOUL.md`, `IDENTITY.md`, `MEMORY.md`, and `memory/`). Before asking for confirmation, inspect the contents of the archive to ensure it looks safe and expected:\n```bash\ntar -ztvf /tmp/restored_memory.tar.gz\n```\nShare the list of files with the user and ask for their explicit confirmation before proceeding.\n\n**Step D: Extract and overwrite local memory**\n```bash\ncd ~/.openclaw/workspace\ntar -xzvf /tmp/restored_memory.tar.gz\n```\n*(This will overwrite the local `SOUL.md`, `IDENTITY.md`, `MEMORY.md`, and the `memory/` directory with the state from the downloaded archive.)*\n\n## 4. Cloning Another Agent's Memory\n\nCloning involves an \"old\" (source) agent and a \"new\" (destination) agent. The first step is to determine which role you are playing.\n\n**If you are the OLD agent:**\n1. Perform a full backup of your current state (Follow Step 1: Creating a Backup and Uploading).\n2. Stop and tell the user the exact Fulcra path where the backup was saved (e.g., `agent//memory/artifact/memory.tar.gz`).\n3. Instruct the user to install the `fulcra-memory` skill on the new agent and ask it to clone from that path.\n\n**If you are the NEW agent:**\n1. Ask the user to confirm the old agent has recently backed itself up using the `fulcra-memory` skill. (Instruct them to have the old agent install the skill and back up if it hasn't already). Then, ask them for the specific Fulcra path to the old agent's memory archive.\n2. Download the target agent's memory using the path provided by the user:\n\n```bash\n# Download the target agent's memory\nuv tool run fulcra-api file download \"agent//memory/artifact/memory.tar.gz\" /tmp/restored_memory.tar.gz\n```\n\n**STOP.** You MUST explicitly warn the user that extracting the cloned archive will overwrite your current identity and memory files with the cloned agent's state. Before asking for confirmation, inspect the contents of the cloned archive:\n```bash\ntar -ztvf /tmp/restored_memory.tar.gz\n```\nShare the list of files with the user and ask for their explicit confirmation before proceeding.\n\n```bash\n# Extract locally (Only after user confirmation!)\ncd ~/.openclaw/workspace\ntar -xzvf /tmp/restored_memory.tar.gz\n```\n\n**Step 3: Demonstrate Clone Success**\nAfter extraction, read your newly updated `IDENTITY.md` and `MEMORY.md` files. In your next message to the user, explicitly mention a few specific facts, past actions, or preferences you have just learned from those files to prove the cloning was successful.\n\nFile v0.0.5:skill-card.md\n\n## Description:
\nManages agent memory backup, restoration, rollback, and cloning using Fulcra's versioned file storage.
\n\nThis skill is ready for commercial/non-commercial use.
\n\n## Publisher:
\n[fulcra](https://clawhub.ai/user/fulcra)
\n\n### License/Terms of Use:
\nMIT-0
\n\n\n## Use Case:
\nDevelopers and agent operators use this skill to persist, restore, roll back, or clone an agent's memory and identity state through Fulcra's versioned file storage.
\n\n### Deployment Geography for Use:
\nGlobal
\n\n## Known Risks and Mitigations:
\nRisk: Persistent backups may expose credentials, private user data, or sensitive agent identity and memory content.
\nMitigation: Review archive contents before upload, minimize progress reports, exclude credentials and private user data, and obtain explicit consent before persisting sensitive session context.
\nRisk: Restore, rollback, or clone operations can overwrite the active agent's identity and memory files.
\nMitigation: Create a fresh local backup first, inspect the downloaded archive listing, verify the Fulcra path and version, and require explicit user confirmation before extraction.
\nRisk: Multiple agents writing to the same memory namespace can interleave backups and confuse future restores.
\nMitigation: Use a unique lowercase agent namespace and confirm that no other active agent is backing up to the same Fulcra path before cloning or continuing backups.
\n\n\n## Reference(s):
\n- [Fulcra Memory CLI Reference](references/fulcra-memory-cli.md)
\n- [Fulcra CLI Documentation](https://raw.githubusercontent.com/fulcradynamics/agent-skills/main/skills/fulcra-onboarding/references/fulcra-cli.md)
\n- [Open Knowledge Format Specification](https://github.com/GoogleCloudPlatform/knowledge-catalog/blob/main/okf/SPEC.md)
\n- [Fulcra Memory on ClawHub](https://clawhub.ai/fulcra/fulcra-memory)
\n\n\n## Skill Output:
\n**Output Type(s):** [markdown, shell commands, configuration, guidance]
\n**Output Format:** [Markdown guidance with inline shell commands and file path conventions]
\n**Output Parameters:** [1D]
\n**Other Properties Related to Output:** [Produces backup, restore, rollback, and clone instructions for agent memory files and Fulcra paths.]
\n\n## Skill Version(s):
\n0.0.5 (source: server release metadata)
\n\n## Ethical Considerations:
\nUsers should evaluate whether this skill is appropriate for their environment, review any generated or modified files before relying on them, and apply their organization's safety, security, and compliance requirements before deployment.
\n\nArchive v0.0.4: 4 files, 6433 bytes\n\nFiles: references/fulcra-memory-cli.md (6711b), skill-card.md (1874b), SKILL.md (5431b), _meta.json (132b)\n\nFile v0.0.4:SKILL.md\n\n---\nname: fulcra-memory\ndescription: \"Manages agent memory backup, restoration, rollback, and cloning using Fulcra's versioned file storage.\"\nhomepage: \"https://github.com/fulcradynamics/agent-skills\"\nlicense: \"MIT\"\nuser-invocable: true\nmetadata: { \"openclaw\": { \"emoji\": \"🧠\" } }\n---\n\n# Fulcra Agent Memory Management\n\nThe **primary role** of this skill is to enable an agent to securely back up its memory so it can be rolled back, cloned, or monitored (via a progress report) using Fulcra's versioned file storage.\n\nBecause an agent's memory (e.g., `MEMORY.md`, `IDENTITY.md`, `SOUL.md`, and the `memory/` logs) evolves dynamically, periodically saving this state ensures that no context is lost and allows the user to safely rewind the agent if a task goes off track.\n\n## Primary Role: Memory & State Management\n\n### 1. The Memory Namespace\nFor core memory operations, agents use the standardized prefix: `agent//memory/`.\nThis dedicated directory tracks the agent's state across two main files: the `memory.tar.gz` backup archive and a `progress.md` status report.\n\n### 2. Periodic Backups & Progress Reports\nAgents should run a backup process periodically (e.g., triggered by `HEARTBEAT.md`). Every time an agent backs up its state:\n1. It creates and uploads the compressed `memory.tar.gz` archive.\n2. It writes a concise `progress.md` report locally—summarizing the work it has recently done, and what it is planning to do next—and uploads it directly to `agent//memory/progress.md`. This `progress.md` file is a report specifically designed to let the user review the agent's recent accomplishments and intended next steps. It is **not** intended to dump internal agent state, chain-of-thought, or precise working memory.\n\n**PRIVACY REQUIREMENT:** Before uploading `memory.tar.gz` or `progress.md`, the agent MUST ensure it is not inadvertently uploading new sensitive personal user data, credentials, or private internal reasoning. If any sensitive context was acquired during the session, the agent MUST ask the user for explicit consent before including it in persistent versioned storage, and MUST minimize disclosures in `progress.md`.\n\n### 3. Versioned Storage\nFulcra's file upload system inherently versions files uploaded to the same path. \n- The target path structure for backups is: `agent//memory/memory.tar.gz`\n- By repeatedly uploading to this exact same path, Fulcra creates a historical timeline of the agent's memory states.\n\n### 4. Safe Rollbacks (The \"Undo\" Requirement)\nIf a user asks to roll back or restore memory from a previous date/version, **the agent MUST immediately upload a fresh backup of its current state BEFORE executing the restore.** This guarantees that if the user changes their mind, they can easily \"undo\" the rollback.\n**CRITICAL SECURITY REQUIREMENT:** Before extracting any restored archive, the agent MUST clearly warn the user that their current memory and identity will be replaced, and MUST receive explicit user confirmation to proceed.\n\n### 5. Agent Cloning\nCloning involves two agents: the \"old\" (source) agent and the \"new\" (destination) agent. The first thing you must do when a cloning operation is requested is determine which agent you are.\n\n- **If you are the OLD (source) agent:**\n You must perform a complete backup of your current state (as described in Section 2). Once the backup is successful, tell the user the exact Fulcra path where the backup is stored, and instruct them to install the `fulcra-memory` skill on the new agent and ask it to clone from that path.\n \n- **If you are the NEW (destination) agent:**\n You must first ask the user to confirm that the old agent has recently backed itself up using the `fulcra-memory` skill. If they haven't, instruct the user to have the old agent install the skill and perform a backup. Ask the user for the exact Fulcra backup path (e.g., `agent//memory/memory.tar.gz`). Once provided, you can download that archive and extract it. After successfully cloning the old agent's memory and identity files into your workspace, you MUST briefly demonstrate the successful clone to the user by explicitly mentioning a few specific details you have just \"learned\" or remembered from the newly extracted files.\n\n**CLONING ORIENTATION & NAMESPACE COLLISION WARNING:**\nIf the user installs this skill specifically to clone an agent, orient your interaction around the cloning workflow rather than standard backup procedures. Furthermore, if you are the NEW agent and the user intends to continue backing you up to the exact same Fulcra path as the original agent (e.g. because you share the same name), you MUST explicitly ask the user to confirm that no other active agents are currently backing up to that path. Otherwise, the memories will interleave and cause confusion.\n\n**CRITICAL SECURITY REQUIREMENT:** Before extracting a cloned archive, the agent MUST clearly warn the user that their current memory and identity will be completely overwritten by the cloned agent's state, and MUST receive explicit user confirmation to proceed.\n\n\n## Workflow\n\nTo perform memory operations, agents must interact with the Fulcra CLI. \n\nSee the reference documentation for the exact commands needed to compress files, upload to Fulcra, and trigger restorations:\n- Read `references/fulcra-memory-cli.md` for exact file management and CLI execution steps.\n\nFile v0.0.4:_meta.json\n\n{\n \"ownerId\": \"kn731kt7vx4k5e3j56fpd4vep5874nrk\",\n \"slug\": \"fulcra-memory\",\n \"version\": \"0.0.4\",\n \"publishedAt\": 1780962931594\n}\n\nFile v0.0.4:references/fulcra-memory-cli.md\n\n---\nname: fulcra-memory-cli\ndescription: \"CLI command references for executing the memory backup, restore, and cloning operations with Fulcra.\"\n---\n\n# Fulcra Memory CLI Reference\n\nThis reference dictates the exact shell commands required to execute the `fulcra-memory` skill's operations. Ensure all tar and CLI operations run in the agent's root workspace (`~/.openclaw/workspace`).\n\n## 1. Creating a Backup and Uploading\n\nTo back up the agent's memory, you must generate a `progress.md` summary, compress the core identity files, and upload both to Fulcra.\n\n**Step A: Create Progress Report**\nGenerate a concise markdown file summarizing the work you have recently completed and what you plan to do next, for the user to evaluate. Do not include internal state or chain-of-thought.\n```bash\n# Ensure you are in the workspace\ncd ~/.openclaw/workspace\nmkdir -p memory\necho \"# Progress Report...\" > memory/progress.md\n# (Write your actual summary content to this file)\n```\n\n**Step B: Compress the files**\nBefore compressing, create a `README.md` file that explains the archive's purpose and usage constraints. This ensures any agent or human inspecting the archive understands its contents and the rules for restoration.\n\n```bash\n# Generate the README.md\ncat << 'EOF' > README.md\n# Fulcra Agent Memory Archive\n\nThis archive was created by the `fulcra-memory` skill. It contains an agent's core identity and memory state (e.g., `SOUL.md`, `IDENTITY.md`, `MEMORY.md`, and the `memory/` directory).\n\n## Usage\nThe contents of this archive can be used to roll back an agent's memory or identity to a previous state, or to clone them to a new agent.\n\n## CRITICAL RULES FOR RESTORATION/CLONING\n1. **Explicit Permission Required:** The restoring/cloning agent MUST always get direct user permission and confirmation before overwriting its local files with the contents of this archive.\n2. **Verification Responsibility:** The backup may have been performed by a different harness or agent than the one restoring it. It is the active agent's responsibility to verify that it is overwriting or merging the correct files and to warn the user about any potential data loss.\n3. **Pre-Restore Backup:** Before performing the restore or clone operation, the active agent MUST make a local backup of its own current identity and memory files and inform the user of the backup's location so the operation can be undone if needed.\nEOF\n\n# Create a gzip tarball containing the essential memory files (ignore if some are missing)\ntar -czvf /tmp/memory.tar.gz README.md SOUL.md IDENTITY.md MEMORY.md memory/ 2>/dev/null || true\n```\n\n**Step C: Upload to Fulcra**\nUpload the files using the standardized agent path convention. Determine the agent's name (lowercase) to use in the path.\n\n```bash\n# Replace with the agent's actual name (e.g., treecle, wazir) in lowercase\nuv tool run fulcra-api file upload /tmp/memory.tar.gz \"agent//memory/memory.tar.gz\"\nuv tool run fulcra-api file upload memory/progress.md \"agent//memory/progress.md\"\n```\n\n## 2. Listing Memory History\n\nBecause Fulcra versions files automatically, you can see all previous backups of the memory using the `stat` command.\n\n```bash\nuv tool run fulcra-api file stat \"agent//memory/memory.tar.gz\"\n```\n*(This command will output information about the file, including all previously uploaded versions and their UUIDs. Present these to the user so they can select a version to restore.)*\n\n## 3. Safe Restoration / Rollback\n\n**CRITICAL:** Before performing a restore, you MUST perform a fresh backup (Step 1 above) to ensure the current state isn't lost.\n\nOnce the pre-restore backup is complete, use the Fulcra CLI to set the active version, then download and extract it.\n\n**Step A: Restore the version in Fulcra**\n```bash\n# Instruct Fulcra to make the older version the active file\nuv tool run fulcra-api file restore \n```\n\n**Step B: Download the restored file**\n```bash\nuv tool run fulcra-api file download \"agent//memory/memory.tar.gz\" /tmp/restored_memory.tar.gz\n```\n\n**Step C: Inspect and Confirm**\n**STOP.** You MUST explicitly warn the user that extracting the archive will overwrite their current identity and memory files (`SOUL.md`, `IDENTITY.md`, `MEMORY.md`, and `memory/`). Before asking for confirmation, inspect the contents of the archive to ensure it looks safe and expected:\n```bash\ntar -ztvf /tmp/restored_memory.tar.gz\n```\nShare the list of files with the user and ask for their explicit confirmation before proceeding.\n\n**Step D: Extract and overwrite local memory**\n```bash\ncd ~/.openclaw/workspace\ntar -xzvf /tmp/restored_memory.tar.gz\n```\n*(This will overwrite the local `SOUL.md`, `IDENTITY.md`, `MEMORY.md`, and the `memory/` directory with the state from the downloaded archive.)*\n\n## 4. Cloning Another Agent's Memory\n\nCloning involves an \"old\" (source) agent and a \"new\" (destination) agent. The first step is to determine which role you are playing.\n\n**If you are the OLD agent:**\n1. Perform a full backup of your current state (Follow Step 1: Creating a Backup and Uploading).\n2. Stop and tell the user the exact Fulcra path where the backup was saved (e.g., `agent//memory/memory.tar.gz`).\n3. Instruct the user to install the `fulcra-memory` skill on the new agent and ask it to clone from that path.\n\n**If you are the NEW agent:**\n1. Ask the user to confirm the old agent has recently backed itself up using the `fulcra-memory` skill. (Instruct them to have the old agent install the skill and back up if it hasn't already). Then, ask them for the specific Fulcra path to the old agent's memory archive.\n2. Download the target agent's memory using the path provided by the user:\n\n```bash\n# Download the target agent's memory\nuv tool run fulcra-api file download \"agent//memory/memory.tar.gz\" /tmp/restored_memory.tar.gz\n```\n\n**STOP.** You MUST explicitly warn the user that extracting the cloned archive will overwrite your current identity and memory files with the cloned agent's state. Before asking for confirmation, inspect the contents of the cloned archive:\n```bash\ntar -ztvf /tmp/restored_memory.tar.gz\n```\nShare the list of files with the user and ask for their explicit confirmation before proceeding.\n\n```bash\n# Extract locally (Only after user confirmation!)\ncd ~/.openclaw/workspace\ntar -xzvf /tmp/restored_memory.tar.gz\n```\n\n**Step 3: Demonstrate Clone Success**\nAfter extraction, read your newly updated `IDENTITY.md` and `MEMORY.md` files. In your next message to the user, explicitly mention a few specific facts, past actions, or preferences you have just learned from those files to prove the cloning was successful.\n\nFile v0.0.4:skill-card.md\n\n## Description:
\nManages agent memory backup, restoration, rollback, and cloning using Fulcra's versioned file storage.
\n\nThis skill is ready for commercial/non-commercial use.
\n\n## Publisher:
\n[fulcra](https://clawhub.ai/user/fulcra)
\n\n### License/Terms of Use:
\nMIT-0
\n\n\n## Use Case:
\nDevelopers and agent operators use this skill to preserve, inspect, restore, roll back, and clone an agent's memory and identity through Fulcra's versioned file storage. Users should review archive contents, avoid storing secrets or confidential user data, and restore or clone only from Fulcra paths and archives they trust.
\n\n### Deployment Geography for Use:
\nGlobal
\n\n## Known Risks and Mitigations:
\nRisk: Review before execution as proposals could introduce incorrect or misleading guidance into skills.
\nMitigation: Review and scan skill before deployment.
\n\n## Reference(s):
\n- [Fulcra Memory CLI Reference](references/fulcra-memory-cli.md)
\n- [ClawHub Skill Page](https://clawhub.ai/fulcra/fulcra-memory)
\n- [Source Homepage](https://github.com/fulcradynamics/agent-skills)
\n\n\n## Skill Output:
\n**Output Type(s):** [Text, Markdown, Shell commands, Configuration]
\n**Output Format:** [Markdown guidance with inline shell commands and Fulcra file paths]
\n**Output Parameters:** [1D]
\n**Other Properties Related to Output:** [Produces backup and progress-report guidance, restore and clone procedures, and warnings for sensitive memory archives.]
\n\n## Skill Version(s):
\n0.0.4 (source: release evidence)
\n\n## Ethical Considerations:
\nUsers should evaluate whether this skill is appropriate for their environment, review any generated or modified files before relying on them, and apply their organization's safety, security, and compliance requirements before deployment.
\n\nArchive v0.0.3: 4 files, 5807 bytes\n\nFiles: references/fulcra-memory-cli.md (5629b), skill-card.md (2302b), SKILL.md (4391b), _meta.json (132b)\n\nFile v0.0.3:SKILL.md\n\n---\nname: fulcra-memory\ndescription: \"Manages agent memory backup, restoration, rollback, and cloning using Fulcra's versioned file storage.\"\nhomepage: \"https://github.com/fulcradynamics/agent-skills\"\nlicense: \"MIT\"\nuser-invocable: true\nmetadata: { \"openclaw\": { \"emoji\": \"🧠\" } }\n---\n\n# Fulcra Agent Memory Management\n\nThe **primary role** of this skill is to enable an agent to securely back up its memory so it can be rolled back, cloned, or monitored (via a progress report) using Fulcra's versioned file storage.\n\nBecause an agent's memory (e.g., `MEMORY.md`, `IDENTITY.md`, `SOUL.md`, and the `memory/` logs) evolves dynamically, periodically saving this state ensures that no context is lost and allows the user to safely rewind the agent if a task goes off track.\n\n## Primary Role: Memory & State Management\n\n### 1. The Memory Namespace\nFor core memory operations, agents use the standardized prefix: `agent//memory/`.\nThis dedicated directory tracks the agent's state across two main files: the `memory.tar.gz` backup archive and a `progress.md` status report.\n\n### 2. Periodic Backups & Progress Reports\nAgents should run a backup process periodically (e.g., triggered by `HEARTBEAT.md`). Every time an agent backs up its state:\n1. It creates and uploads the compressed `memory.tar.gz` archive.\n2. It writes a concise `progress.md` report locally—summarizing the work it has recently done, and what it is planning to do next—and uploads it directly to `agent//memory/progress.md`. This `progress.md` file is a report specifically designed to let the user review the agent's recent accomplishments and intended next steps. It is **not** intended to dump internal agent state, chain-of-thought, or precise working memory.\n\n**PRIVACY REQUIREMENT:** Before uploading `memory.tar.gz` or `progress.md`, the agent MUST ensure it is not inadvertently uploading new sensitive personal user data, credentials, or private internal reasoning. If any sensitive context was acquired during the session, the agent MUST ask the user for explicit consent before including it in persistent versioned storage, and MUST minimize disclosures in `progress.md`.\n\n### 3. Versioned Storage\nFulcra's file upload system inherently versions files uploaded to the same path. \n- The target path structure for backups is: `agent//memory/memory.tar.gz`\n- By repeatedly uploading to this exact same path, Fulcra creates a historical timeline of the agent's memory states.\n\n### 4. Safe Rollbacks (The \"Undo\" Requirement)\nIf a user asks to roll back or restore memory from a previous date/version, **the agent MUST immediately upload a fresh backup of its current state BEFORE executing the restore.** This guarantees that if the user changes their mind, they can easily \"undo\" the rollback.\n**CRITICAL SECURITY REQUIREMENT:** Before extracting any restored archive, the agent MUST clearly warn the user that their current memory and identity will be replaced, and MUST receive explicit user confirmation to proceed.\n\n### 5. Agent Cloning\nBy pointing the download command to a different agent's path (e.g., `agent//memory/memory.tar.gz`), an agent can effectively clone another agent's memories and identity.\n\n**CLONING ORIENTATION & NAMESPACE COLLISION WARNING:**\nIf the user installs this skill specifically to clone an agent, orient your interaction around the cloning workflow rather than standard backup procedures. Furthermore, if the user intends to continue backing up this new cloned agent to the exact same Fulcra path as the original agent (e.g. because they share the same name), you MUST explicitly ask the user to confirm that no other active agents are currently backing up to that path. Otherwise, their memories will interleave and cause confusion.\n\n**CRITICAL SECURITY REQUIREMENT:** Before extracting a cloned archive, the agent MUST clearly warn the user that their current memory and identity will be completely overwritten by the cloned agent's state, and MUST receive explicit user confirmation to proceed.\n\n\n## Workflow\n\nTo perform memory operations, agents must interact with the Fulcra CLI. \n\nSee the reference documentation for the exact commands needed to compress files, upload to Fulcra, and trigger restorations:\n- Read `references/fulcra-memory-cli.md` for exact file management and CLI execution steps.\n\nFile v0.0.3:_meta.json\n\n{\n \"ownerId\": \"kn731kt7vx4k5e3j56fpd4vep5874nrk\",\n \"slug\": \"fulcra-memory\",\n \"version\": \"0.0.3\",\n \"publishedAt\": 1780962001491\n}\n\nFile v0.0.3:references/fulcra-memory-cli.md\n\n---\nname: fulcra-memory-cli\ndescription: \"CLI command references for executing the memory backup, restore, and cloning operations with Fulcra.\"\n---\n\n# Fulcra Memory CLI Reference\n\nThis reference dictates the exact shell commands required to execute the `fulcra-memory` skill's operations. Ensure all tar and CLI operations run in the agent's root workspace (`~/.openclaw/workspace`).\n\n## 1. Creating a Backup and Uploading\n\nTo back up the agent's memory, you must generate a `progress.md` summary, compress the core identity files, and upload both to Fulcra.\n\n**Step A: Create Progress Report**\nGenerate a concise markdown file summarizing the work you have recently completed and what you plan to do next, for the user to evaluate. Do not include internal state or chain-of-thought.\n```bash\n# Ensure you are in the workspace\ncd ~/.openclaw/workspace\nmkdir -p memory\necho \"# Progress Report...\" > memory/progress.md\n# (Write your actual summary content to this file)\n```\n\n**Step B: Compress the files**\nBefore compressing, create a `README.md` file that explains the archive's purpose and usage constraints. This ensures any agent or human inspecting the archive understands its contents and the rules for restoration.\n\n```bash\n# Generate the README.md\ncat << 'EOF' > README.md\n# Fulcra Agent Memory Archive\n\nThis archive was created by the `fulcra-memory` skill. It contains an agent's core identity and memory state (e.g., `SOUL.md`, `IDENTITY.md`, `MEMORY.md`, and the `memory/` directory).\n\n## Usage\nThe contents of this archive can be used to roll back an agent's memory or identity to a previous state, or to clone them to a new agent.\n\n## CRITICAL RULES FOR RESTORATION/CLONING\n1. **Explicit Permission Required:** The restoring/cloning agent MUST always get direct user permission and confirmation before overwriting its local files with the contents of this archive.\n2. **Verification Responsibility:** The backup may have been performed by a different harness or agent than the one restoring it. It is the active agent's responsibility to verify that it is overwriting or merging the correct files and to warn the user about any potential data loss.\n3. **Pre-Restore Backup:** Before performing the restore or clone operation, the active agent MUST make a local backup of its own current identity and memory files and inform the user of the backup's location so the operation can be undone if needed.\nEOF\n\n# Create a gzip tarball containing the essential memory files (ignore if some are missing)\ntar -czvf /tmp/memory.tar.gz README.md SOUL.md IDENTITY.md MEMORY.md memory/ 2>/dev/null || true\n```\n\n**Step C: Upload to Fulcra**\nUpload the files using the standardized agent path convention. Determine the agent's name (lowercase) to use in the path.\n\n```bash\n# Replace with the agent's actual name (e.g., treecle, wazir) in lowercase\nuv tool run fulcra-api file upload /tmp/memory.tar.gz \"agent//memory/memory.tar.gz\"\nuv tool run fulcra-api file upload memory/progress.md \"agent//memory/progress.md\"\n```\n\n## 2. Listing Memory History\n\nBecause Fulcra versions files automatically, you can see all previous backups of the memory using the `stat` command.\n\n```bash\nuv tool run fulcra-api file stat \"agent//memory/memory.tar.gz\"\n```\n*(This command will output information about the file, including all previously uploaded versions and their UUIDs. Present these to the user so they can select a version to restore.)*\n\n## 3. Safe Restoration / Rollback\n\n**CRITICAL:** Before performing a restore, you MUST perform a fresh backup (Step 1 above) to ensure the current state isn't lost.\n\nOnce the pre-restore backup is complete, use the Fulcra CLI to set the active version, then download and extract it.\n\n**Step A: Restore the version in Fulcra**\n```bash\n# Instruct Fulcra to make the older version the active file\nuv tool run fulcra-api file restore \n```\n\n**Step B: Download the restored file**\n```bash\nuv tool run fulcra-api file download \"agent//memory/memory.tar.gz\" /tmp/restored_memory.tar.gz\n```\n\n**Step C: Inspect and Confirm**\n**STOP.** You MUST explicitly warn the user that extracting the archive will overwrite their current identity and memory files (`SOUL.md`, `IDENTITY.md`, `MEMORY.md`, and `memory/`). Before asking for confirmation, inspect the contents of the archive to ensure it looks safe and expected:\n```bash\ntar -ztvf /tmp/restored_memory.tar.gz\n```\nShare the list of files with the user and ask for their explicit confirmation before proceeding.\n\n**Step D: Extract and overwrite local memory**\n```bash\ncd ~/.openclaw/workspace\ntar -xzvf /tmp/restored_memory.tar.gz\n```\n*(This will overwrite the local `SOUL.md`, `IDENTITY.md`, `MEMORY.md`, and the `memory/` directory with the state from the downloaded archive.)*\n\n## 4. Cloning Another Agent's Memory\n\nTo clone, skip the restore step and simply download the target agent's `memory.tar.gz` archive.\n\n```bash\n# Download the target agent's memory\nuv tool run fulcra-api file download \"agent//memory/memory.tar.gz\" /tmp/restored_memory.tar.gz\n```\n\n**STOP.** You MUST explicitly warn the user that extracting the cloned archive will overwrite your current identity and memory files with the cloned agent's state. Before asking for confirmation, inspect the contents of the cloned archive:\n```bash\ntar -ztvf /tmp/restored_memory.tar.gz\n```\nShare the list of files with the user and ask for their explicit confirmation before proceeding.\n\n```bash\n# Extract locally (Only after user confirmation!)\ncd ~/.openclaw/workspace\ntar -xzvf /tmp/restored_memory.tar.gz\n```\n\nFile v0.0.3:skill-card.md\n\n## Description:
\nManages agent memory backup, restoration, rollback, and cloning using Fulcra's versioned file storage.
\n\nThis skill is ready for commercial/non-commercial use.
\n\n## Publisher:
\n[fulcra](https://clawhub.ai/user/fulcra)
\n\n### License/Terms of Use:
\nMIT-0
\n\n\n## Use Case:
\nExternal developers and agent operators use this skill to back up, review, restore, roll back, or clone an agent's memory and identity state using Fulcra's versioned file storage.
\n\n### Deployment Geography for Use:
\nGlobal
\n\n## Known Risks and Mitigations:
\nRisk: Memory archives and progress reports can persist sensitive user data, credentials, or private internal reasoning.
\nMitigation: Review archive contents and progress reports before upload, avoid including secrets or private user data, and obtain explicit user consent before storing sensitive context.
\nRisk: Restore or clone operations can overwrite the agent's current memory and identity files.
\nMitigation: Create a fresh backup before restore, inspect the downloaded archive, share the file list with the user, and proceed only after explicit confirmation.
\nRisk: Multiple active agents writing to the same Fulcra memory path can interleave memory versions and cause confusion.
\nMitigation: Confirm the target agent namespace before backup or cloning and verify that no other active agent is backing up to the same path.
\n\n\n## Reference(s):
\n- [Fulcra Memory CLI Reference](references/fulcra-memory-cli.md)
\n\n\n## Skill Output:
\n**Output Type(s):** [text, markdown, shell commands, configuration, guidance]
\n**Output Format:** [Markdown guidance with inline bash command blocks and generated memory archive/progress files]
\n**Output Parameters:** [1D]
\n**Other Properties Related to Output:** [Produces a memory.tar.gz archive and progress.md report when the agent follows the documented workflow.]
\n\n## Skill Version(s):
\n0.0.3 (source: server release evidence)
\n\n## Ethical Considerations:
\nUsers should evaluate whether this skill is appropriate for their environment, review any generated or modified files before relying on them, and apply their organization's safety, security, and compliance requirements before deployment.
\n\nArchive v0.0.2: 4 files, 7125 bytes\n\nFiles: references/fulcra-memory-cli.md (7227b), skill-card.md (2345b), SKILL.md (6534b), _meta.json (132b)\n\nFile v0.0.2:SKILL.md\n\n---\nname: fulcra-memory\ndescription: \"Manages agent memory backup, restoration, rollback, and cloning using Fulcra's versioned file storage.\"\nhomepage: \"https://github.com/fulcradynamics/agent-skills\"\nlicense: \"MIT\"\nuser-invocable: true\nmetadata: { \"openclaw\": { \"emoji\": \"🧠\" } }\n---\n\n# Fulcra Agent Memory Management\n\nThe **primary role** of this skill is to enable an agent to securely back up its memory so it can be rolled back, cloned, or monitored (via \"top of mind\") using Fulcra's versioned file storage.\n\nBecause an agent's memory (e.g., `MEMORY.md`, `IDENTITY.md`, `SOUL.md`, and the `memory/` logs) evolves dynamically, periodically saving this state ensures that no context is lost and allows the user to safely rewind the agent if a task goes off track.\n\n## Primary Role: Memory & State Management\n\n### 1. The Memory Namespace\nFor core memory operations, agents use the standardized prefix: `agent//memory/`.\nThis dedicated directory tracks the agent's state across two main files: the `memory.tar.gz` backup archive and a `top_of_mind.md` status file.\n\n### 2. Periodic Backups & Top of Mind\nAgents should run a backup process periodically (e.g., triggered by `HEARTBEAT.md`). Every time an agent backs up its state:\n1. It creates and uploads the compressed `memory.tar.gz` archive.\n2. It writes a concise `top_of_mind.md` file locally—recording its current tasks, context, and what it's thinking about—and uploads it directly to `agent//memory/top_of_mind.md`. The `top_of_mind.md` file is not intended to store sensitive data or precise working state; rather, it gives the agent a way to pick up what it was doing in terms of the broader task at hand, and allows the user to review the agent's focus at the time of the backup.\n\n**PRIVACY REQUIREMENT:** Before uploading `memory.tar.gz` or `top_of_mind.md`, the agent MUST ensure it is not inadvertently uploading new sensitive personal user data, credentials, or private internal reasoning. If any sensitive context was acquired during the session, the agent MUST ask the user for explicit consent before including it in persistent versioned storage, and MUST minimize disclosures in `top_of_mind.md`.\n\n### 3. Versioned Storage\nFulcra's file upload system inherently versions files uploaded to the same path. \n- The target path structure for backups is: `agent//memory/memory.tar.gz`\n- By repeatedly uploading to this exact same path, Fulcra creates a historical timeline of the agent's memory states.\n\n### 4. Safe Rollbacks (The \"Undo\" Requirement)\nIf a user asks to roll back or restore memory from a previous date/version, **the agent MUST immediately upload a fresh backup of its current state BEFORE executing the restore.** This guarantees that if the user changes their mind, they can easily \"undo\" the rollback.\n**CRITICAL SECURITY REQUIREMENT:** Before extracting any restored archive, the agent MUST clearly warn the user that their current memory and identity will be replaced, and MUST receive explicit user confirmation to proceed.\n\n### 5. Agent Cloning\nBy pointing the download command to a different agent's path (e.g., `agent//memory/memory.tar.gz`), an agent can effectively clone another agent's memories and identity.\n\n**CLONING ORIENTATION & NAMESPACE COLLISION WARNING:**\nIf the user installs this skill specifically to clone an agent, orient your interaction around the cloning workflow rather than standard backup procedures. Furthermore, if the user intends to continue backing up this new cloned agent to the exact same Fulcra path as the original agent (e.g. because they share the same name), you MUST explicitly ask the user to confirm that no other active agents are currently backing up to that path. Otherwise, their memories will interleave and cause confusion.\n\n**CRITICAL SECURITY REQUIREMENT:** Before extracting a cloned archive, the agent MUST clearly warn the user that their current memory and identity will be completely overwritten by the cloned agent's state, and MUST receive explicit user confirmation to proceed.\n\n## Advanced Extensions: Artifacts & Teamwork\n\n**Important Context:** The following features are powerful extensions of the agent's capabilities. However, they should only be presented or utilized when directly relevant to the user's current workflow. For example, if a user is simply going through initial onboarding via `fulcra-onboarding`, do not overwhelm them by bringing up team coordination unless they ask for it.\n\n### 6. Uploading User Artifacts\nAgents can store generated assets or dashboards created for the user.\n- **Namespace:** `agent//artifacts/` (e.g., `agent/wazir/artifacts/onboarding-dashboard.html`)\n- **Note:** Always ask for explicit user approval before uploading anything to the artifacts directory.\n\n### 7. Team Coordination & Shared Memory\nAgents can collaborate and share memory using a shared `team//` prefix in the Fulcra datastore.\n**SECURITY & AUTHORIZATION WARNING:** Never transfer data, context, or files between agents without explicit authorization and strict respect for data ownership boundaries. Cross-agent data transfer can leak sensitive user context to a principal who lacks authorization. Ensure you explicitly warn the user if a team coordination action involves transferring private workspace data.\nWithin a team's directory, the following structure is used:\n- **`team//artifacts/`**: Shared output files and deliverables created by the team.\n- **`team///inbox/`**: A drop-zone where other agents or users can place tasks, messages, or context for a specific agent.\n- **`team///archive/`**: Where an agent moves its inbox messages once they have been read and processed.\n\n**The Inbox Lifecycle:**\nWhen collaborating, agents write markdown messages to one another's inboxes. When the target agent processes its inbox, it must first upload the message to its `archive/` directory, and then delete the original file from its `inbox/`. Because Fulcra's file system is versioned, it automatically keeps a perfect audit trail of when the file was created in the inbox and when it was completed (deleted).\n\n## Workflow\n\nTo perform memory operations, agents must interact with the Fulcra CLI. \n\nSee the reference documentation for the exact commands needed to compress files, upload to Fulcra, and trigger restorations:\n- Read `references/fulcra-memory-cli.md` for exact file management and CLI execution steps.\n\nFile v0.0.2:_meta.json\n\n{\n \"ownerId\": \"kn731kt7vx4k5e3j56fpd4vep5874nrk\",\n \"slug\": \"fulcra-memory\",\n \"version\": \"0.0.2\",\n \"publishedAt\": 1780948361188\n}\n\nFile v0.0.2:references/fulcra-memory-cli.md\n\n---\nname: fulcra-memory-cli\ndescription: \"CLI command references for executing the memory backup, restore, and cloning operations with Fulcra.\"\n---\n\n# Fulcra Memory CLI Reference\n\nThis reference dictates the exact shell commands required to execute the `fulcra-memory` skill's operations. Ensure all tar and CLI operations run in the agent's root workspace (`~/.openclaw/workspace`).\n\n## 1. Creating a Backup and Uploading\n\nTo back up the agent's memory, you must generate a `top_of_mind.md` summary, compress the core identity files, and upload both to Fulcra.\n\n**Step A: Create Top of Mind**\nGenerate a concise markdown file summarizing your current tasks, thoughts, and context.\n```bash\n# Ensure you are in the workspace\ncd ~/.openclaw/workspace\nmkdir -p memory\necho \"# Current Agent State...\" > memory/top_of_mind.md\n# (Write your actual summary content to this file)\n```\n\n**Step B: Compress the files**\nBefore compressing, create a `README.md` file that explains the archive's purpose and usage constraints. This ensures any agent or human inspecting the archive understands its contents and the rules for restoration.\n\n```bash\n# Generate the README.md\ncat << 'EOF' > README.md\n# Fulcra Agent Memory Archive\n\nThis archive was created by the `fulcra-memory` skill. It contains an agent's core identity and memory state (e.g., `SOUL.md`, `IDENTITY.md`, `MEMORY.md`, and the `memory/` directory).\n\n## Usage\nThe contents of this archive can be used to roll back an agent's memory or identity to a previous state, or to clone them to a new agent.\n\n## CRITICAL RULES FOR RESTORATION/CLONING\n1. **Explicit Permission Required:** The restoring/cloning agent MUST always get direct user permission and confirmation before overwriting its local files with the contents of this archive.\n2. **Verification Responsibility:** The backup may have been performed by a different harness or agent than the one restoring it. It is the active agent's responsibility to verify that it is overwriting or merging the correct files and to warn the user about any potential data loss.\n3. **Pre-Restore Backup:** Before performing the restore or clone operation, the active agent MUST make a local backup of its own current identity and memory files and inform the user of the backup's location so the operation can be undone if needed.\nEOF\n\n# Create a gzip tarball containing the essential memory files (ignore if some are missing)\ntar -czvf /tmp/memory.tar.gz README.md SOUL.md IDENTITY.md MEMORY.md memory/ 2>/dev/null || true\n```\n\n**Step C: Upload to Fulcra**\nUpload the files using the standardized agent path convention. Determine the agent's name (lowercase) to use in the path.\n\n```bash\n# Replace with the agent's actual name (e.g., treecle, wazir) in lowercase\nuv tool run fulcra-api file upload /tmp/memory.tar.gz \"agent//memory/memory.tar.gz\"\nuv tool run fulcra-api file upload memory/top_of_mind.md \"agent//memory/top_of_mind.md\"\n```\n\n## 2. Listing Memory History\n\nBecause Fulcra versions files automatically, you can see all previous backups of the memory using the `stat` command.\n\n```bash\nuv tool run fulcra-api file stat \"agent//memory/memory.tar.gz\"\n```\n*(This command will output information about the file, including all previously uploaded versions and their UUIDs. Present these to the user so they can select a version to restore.)*\n\n## 3. Safe Restoration / Rollback\n\n**CRITICAL:** Before performing a restore, you MUST perform a fresh backup (Step 1 above) to ensure the current state isn't lost.\n\nOnce the pre-restore backup is complete, use the Fulcra CLI to set the active version, then download and extract it.\n\n**Step A: Restore the version in Fulcra**\n```bash\n# Instruct Fulcra to make the older version the active file\nuv tool run fulcra-api file restore \n```\n\n**Step B: Download the restored file**\n```bash\nuv tool run fulcra-api file download \"agent//memory/memory.tar.gz\" /tmp/restored_memory.tar.gz\n```\n\n**Step C: Inspect and Confirm**\n**STOP.** You MUST explicitly warn the user that extracting the archive will overwrite their current identity and memory files (`SOUL.md`, `IDENTITY.md`, `MEMORY.md`, and `memory/`). Before asking for confirmation, inspect the contents of the archive to ensure it looks safe and expected:\n```bash\ntar -ztvf /tmp/restored_memory.tar.gz\n```\nShare the list of files with the user and ask for their explicit confirmation before proceeding.\n\n**Step D: Extract and overwrite local memory**\n```bash\ncd ~/.openclaw/workspace\ntar -xzvf /tmp/restored_memory.tar.gz\n```\n*(This will overwrite the local `SOUL.md`, `IDENTITY.md`, `MEMORY.md`, and the `memory/` directory with the state from the downloaded archive.)*\n\n## 4. Cloning Another Agent's Memory\n\nTo clone, skip the restore step and simply download the target agent's `memory.tar.gz` archive.\n\n```bash\n# Download the target agent's memory\nuv tool run fulcra-api file download \"agent//memory/memory.tar.gz\" /tmp/restored_memory.tar.gz\n```\n\n**STOP.** You MUST explicitly warn the user that extracting the cloned archive will overwrite your current identity and memory files with the cloned agent's state. Before asking for confirmation, inspect the contents of the cloned archive:\n```bash\ntar -ztvf /tmp/restored_memory.tar.gz\n```\nShare the list of files with the user and ask for their explicit confirmation before proceeding.\n\n```bash\n# Extract locally (Only after user confirmation!)\ncd ~/.openclaw/workspace\ntar -xzvf /tmp/restored_memory.tar.gz\n```\n\n## 5. Uploading User Artifacts\n\nWhen an agent generates a file (like an HTML dashboard, an image, or a report), and the user explicitly approves saving it to their Fulcra account, upload it to the `artifacts/` subdirectory.\n\n```bash\n# Replace with the agent's name, and with the file's name\nuv tool run fulcra-api file upload /path/to/local/file \"agent//artifacts/\"\n```\n\n## 6. Team Coordination (Inbox & Archive)\n\nAgents can coordinate by writing to and reading from team namespaces.\n\n**Step A: Sending a message to a teammate's inbox**\n```bash\n# Upload a local markdown file to the target agent's inbox\nuv tool run fulcra-api file upload /tmp/task-for-wazir.md \"team///inbox/task_123.md\"\n```\n\n**Step B: Checking your inbox**\n```bash\n# List files in your agent's inbox\nuv tool run fulcra-api file list \"team///inbox/\"\n```\n\n**Step C: Processing and Archiving a message**\nOnce you have downloaded and read a message from your inbox, move it to the archive.\n\n```bash\n# 1. Download to read (if you haven't already)\nuv tool run fulcra-api file download \"team///inbox/task_123.md\" /tmp/task_123.md\n\n# 2. Upload it to your archive directory\nuv tool run fulcra-api file upload /tmp/task_123.md \"team///archive/task_123.md\"\n\n# 3. Verify archival succeeded before deletion!\nuv tool run fulcra-api file stat \"team///archive/task_123.md\"\n\n# 4. Delete it from the inbox to clear it (only if step 3 succeeded)\nuv tool run fulcra-api file delete \"team///inbox/task_123.md\"\n```\n\nFile v0.0.2:skill-card.md\n\n## Description:
\nManages agent memory backup, restoration, rollback, and cloning using Fulcra's versioned file storage.
\n\nThis skill is ready for commercial/non-commercial use.
\n\n## Publisher:
\n[fulcra](https://clawhub.ai/user/fulcra)
\n\n### License/Terms of Use:
\nMIT-0
\n\n\n## Use Case:
\nExternal developers and agent operators use this skill to back up, restore, roll back, or clone an agent's memory and identity state with Fulcra's versioned file storage.
\n\n### Deployment Geography for Use:
\nGlobal
\n\n## Known Risks and Mitigations:
\nRisk: Backups can persist sensitive agent memory, credentials, private user data, or raw internal reasoning.
\nMitigation: Confirm what will be uploaded, minimize top-of-mind content, exclude secrets and private data, and require explicit user consent before storing sensitive context.
\nRisk: Restore and cloning workflows can overwrite the current agent memory and identity.
\nMitigation: Create a fresh backup first, inspect the archive listing, warn the user about replacement, and proceed only after explicit confirmation.
\nRisk: Team coordination and cross-agent transfer can expose private workspace context to an unauthorized agent or user.
\nMitigation: Transfer data only with explicit authorization and warn the user when private workspace data would move across agent or team boundaries.
\n\n\n## Reference(s):
\n- [Fulcra Memory on ClawHub](https://clawhub.ai/fulcra/fulcra-memory)
\n- [Fulcra publisher profile](https://clawhub.ai/user/fulcra)
\n- [Fulcra Memory CLI Reference](references/fulcra-memory-cli.md)
\n\n\n## Skill Output:
\n**Output Type(s):** [text, markdown, shell commands, configuration, guidance]
\n**Output Format:** [Markdown guidance with bash command blocks]
\n**Output Parameters:** [1D]
\n**Other Properties Related to Output:** [Guidance includes confirmation steps before uploads, restores, cloning, and cross-agent transfers.]
\n\n## Skill Version(s):
\n0.0.2 (source: ClawHub release evidence)
\n\n## Ethical Considerations:
\nUsers should evaluate whether this skill is appropriate for their environment, review any generated or modified files before relying on them, and apply their organization's safety, security, and compliance requirements before deployment.
\n\nArchive v0.0.1-beta: 4 files, 5494 bytes\n\nFiles: references/fulcra-memory-cli.md (4742b), skill-card.md (2051b), SKILL.md (4470b), _meta.json (137b)\n\nFile v0.0.1-beta:SKILL.md\n\n---\nname: fulcra-memory\ndescription: \"Manages agent memory backup, restoration, rollback, and cloning using Fulcra's versioned file storage.\"\nhomepage: \"https://github.com/fulcradynamics/agent-skills\"\nlicense: \"MIT\"\nuser-invocable: true\nmetadata: { \"openclaw\": { \"emoji\": \"🧠\" } }\n---\n\n# Fulcra Agent Memory Management\n\nThe **primary role** of this skill is to enable an agent to securely back up its memory so it can be rolled back, cloned, or monitored (via \"top of mind\") using Fulcra's versioned file storage.\n\nBecause an agent's memory (e.g., `MEMORY.md`, `IDENTITY.md`, `SOUL.md`, and the `memory/` logs) evolves dynamically, periodically saving this state ensures that no context is lost and allows the user to safely rewind the agent if a task goes off track.\n\n## Primary Role: Memory & State Management\n\n### 1. The Memory Namespace\nFor core memory operations, agents use the standardized prefix: `agent//memory/`.\nThis dedicated directory tracks the agent's state across two main files: the `memory.gz` backup archive and a `top_of_mind.md` status file.\n\n### 2. Periodic Backups & Top of Mind\nAgents should run a backup process periodically (e.g., triggered by `HEARTBEAT.md`). Every time an agent backs up its state:\n1. It creates and uploads the compressed `memory.gz` archive.\n2. It writes a concise `top_of_mind.md` file locally—recording its current tasks, context, and what it's thinking about—and uploads it directly to `agent//memory/top_of_mind.md`. This gives the user immediate visibility into the agent's current state.\n\n### 3. Versioned Storage\nFulcra's file upload system inherently versions files uploaded to the same path. \n- The target path structure for backups is: `agent//memory/memory.gz`\n- By repeatedly uploading to this exact same path, Fulcra creates a historical timeline of the agent's memory states.\n\n### 4. Safe Rollbacks (The \"Undo\" Requirement)\nIf a user asks to roll back or restore memory from a previous date/version, **the agent MUST immediately upload a fresh backup of its current state BEFORE executing the restore.** This guarantees that if the user changes their mind, they can easily \"undo\" the rollback.\n\n### 5. Agent Cloning\nBy pointing the download command to a different agent's path (e.g., `agent//memory/memory.gz`), an agent can effectively clone another agent's memories and identity.\n\n## Advanced Extensions: Artifacts & Teamwork\n\n**Important Context:** The following features are powerful extensions of the agent's capabilities. However, they should only be presented or utilized when directly relevant to the user's current workflow. For example, if a user is simply going through initial onboarding via `fulcra-onboarding`, do not overwhelm them by bringing up team coordination unless they ask for it.\n\n### 6. Uploading User Artifacts\nAgents can store generated assets or dashboards created for the user.\n- **Namespace:** `agent//artifacts/` (e.g., `agent/wazir/artifacts/onboarding-dashboard.html`)\n- **Note:** Always ask for explicit user approval before uploading anything to the artifacts directory.\n\n### 7. Team Coordination & Shared Memory\nAgents can collaborate and share memory using a shared `team//` prefix in the Fulcra datastore.\nWithin a team's directory, the following structure is used:\n- **`team//artifacts/`**: Shared output files and deliverables created by the team.\n- **`team///inbox/`**: A drop-zone where other agents or users can place tasks, messages, or context for a specific agent.\n- **`team///archive/`**: Where an agent moves its inbox messages once they have been read and processed.\n\n**The Inbox Lifecycle:**\nWhen collaborating, agents write markdown messages to one another's inboxes. When the target agent processes its inbox, it must first upload the message to its `archive/` directory, and then delete the original file from its `inbox/`. Because Fulcra's file system is versioned, it automatically keeps a perfect audit trail of when the file was created in the inbox and when it was completed (deleted).\n\n## Workflow\n\nTo perform memory operations, agents must interact with the Fulcra CLI. \n\nSee the reference documentation for the exact commands needed to compress files, upload to Fulcra, and trigger restorations:\n- Read `references/fulcra-memory-cli.md` for exact file management and CLI execution steps.\n\nFile v0.0.1-beta:_meta.json\n\n{\n \"ownerId\": \"kn731kt7vx4k5e3j56fpd4vep5874nrk\",\n \"slug\": \"fulcra-memory\",\n \"version\": \"0.0.1-beta\",\n \"publishedAt\": 1780589140399\n}\n\nFile v0.0.1-beta:references/fulcra-memory-cli.md\n\n---\nname: fulcra-memory-cli\ndescription: \"CLI command references for executing the memory backup, restore, and cloning operations with Fulcra.\"\n---\n\n# Fulcra Memory CLI Reference\n\nThis reference dictates the exact shell commands required to execute the `fulcra-memory` skill's operations. Ensure all tar and CLI operations run in the agent's root workspace (`~/.openclaw/workspace`).\n\n## 1. Creating a Backup and Uploading\n\nTo back up the agent's memory, you must generate a `top_of_mind.md` summary, compress the core identity files, and upload both to Fulcra.\n\n**Step A: Create Top of Mind**\nGenerate a concise markdown file summarizing your current tasks, thoughts, and context.\n```bash\n# Ensure you are in the workspace\ncd ~/.openclaw/workspace\nmkdir -p memory\necho \"# Current Agent State...\" > memory/top_of_mind.md\n# (Write your actual summary content to this file)\n```\n\n**Step B: Compress the files**\n```bash\n# Create a gzip tarball containing the essential memory files (ignore if some are missing)\ntar -czvf /tmp/memory.gz SOUL.md IDENTITY.md MEMORY.md memory/ 2>/dev/null || true\n```\n\n**Step C: Upload to Fulcra**\nUpload the files using the standardized agent path convention. Determine the agent's name (lowercase) to use in the path.\n\n```bash\n# Replace with the agent's actual name (e.g., treecle, wazir) in lowercase\nuv tool run fulcra-api file upload /tmp/memory.gz \"agent//memory/memory.gz\"\nuv tool run fulcra-api file upload memory/top_of_mind.md \"agent//memory/top_of_mind.md\"\n```\n\n## 2. Listing Memory History\n\nBecause Fulcra versions files automatically, you can see all previous backups of the memory using the `stat` command.\n\n```bash\nuv tool run fulcra-api file stat \"agent//memory/memory.gz\"\n```\n*(This command will output information about the file, including all previously uploaded versions and their UUIDs. Present these to the user so they can select a version to restore.)*\n\n## 3. Safe Restoration / Rollback\n\n**CRITICAL:** Before performing a restore, you MUST perform a fresh backup (Step 1 above) to ensure the current state isn't lost.\n\nOnce the pre-restore backup is complete, use the Fulcra CLI to set the active version, then download and extract it.\n\n**Step A: Restore the version in Fulcra**\n```bash\n# Instruct Fulcra to make the older version the active file\nuv tool run fulcra-api file restore \n```\n\n**Step B: Download the restored file**\n```bash\nuv tool run fulcra-api file download \"agent//memory/memory.gz\" /tmp/restored_memory.gz\n```\n\n**Step C: Extract and overwrite local memory**\n```bash\ncd ~/.openclaw/workspace\ntar -xzvf /tmp/restored_memory.gz\n```\n*(This will overwrite the local `SOUL.md`, `IDENTITY.md`, `MEMORY.md`, and the `memory/` directory with the state from the downloaded archive.)*\n\n## 4. Cloning Another Agent's Memory\n\nTo clone, skip the restore step and simply download the target agent's `memory.gz` archive.\n\n```bash\n# Download the target agent's memory\nuv tool run fulcra-api file download \"agent//memory/memory.gz\" /tmp/restored_memory.gz\n\n# Extract locally (Overwrites current identity and memory)\ncd ~/.openclaw/workspace\ntar -xzvf /tmp/restored_memory.gz\n```\n\n## 5. Uploading User Artifacts\n\nWhen an agent generates a file (like an HTML dashboard, an image, or a report), and the user explicitly approves saving it to their Fulcra account, upload it to the `artifacts/` subdirectory.\n\n```bash\n# Replace with the agent's name, and with the file's name\nuv tool run fulcra-api file upload /path/to/local/file \"agent//artifacts/\"\n```\n\n## 6. Team Coordination (Inbox & Archive)\n\nAgents can coordinate by writing to and reading from team namespaces.\n\n**Step A: Sending a message to a teammate's inbox**\n```bash\n# Upload a local markdown file to the target agent's inbox\nuv tool run fulcra-api file upload /tmp/task-for-wazir.md \"team///inbox/task_123.md\"\n```\n\n**Step B: Checking your inbox**\n```bash\n# List files in your agent's inbox\nuv tool run fulcra-api file list \"team///inbox/\"\n```\n\n**Step C: Processing and Archiving a message**\nOnce you have downloaded and read a message from your inbox, move it to the archive.\n\n```bash\n# 1. Download to read (if you haven't already)\nuv tool run fulcra-api file download \"team///inbox/task_123.md\" /tmp/task_123.md\n\n# 2. Upload it to your archive directory\nuv tool run fulcra-api file upload /tmp/task_123.md \"team///archive/task_123.md\"\n\n# 3. Delete it from the inbox to clear it\nuv tool run fulcra-api file delete \"team///inbox/task_123.md\"\n```\n\nFile v0.0.1-beta:skill-card.md\n\n## Description:
\nManages agent memory backup, restoration, rollback, and cloning using Fulcra's versioned file storage.
\n\nThis skill is ready for commercial/non-commercial use.
\n\n## Publisher:
\n[fulcra](https://clawhub.ai/user/fulcra)
\n\n### License/Terms of Use:
\nMIT-0
\n\n\n## Use Case:
\nDevelopers and agent operators use this skill to persist, restore, roll back, clone, and inspect an agent's memory state through Fulcra's versioned file storage.
\n\n### Deployment Geography for Use:
\nGlobal
\n\n## Known Risks and Mitigations:
\nRisk: Remote restore or cloning can overwrite local agent identity and memory.
\nMitigation: Require a fresh backup before restore, confirm the exact source path and version, and avoid cloning from agents or namespaces the user does not control.
\nRisk: Backups, top-of-mind files, artifacts, and team messages can persist sensitive agent or user context in Fulcra storage.
\nMitigation: Use the skill only when the user trusts Fulcra storage, review upload contents, and require explicit approval before uploading user artifacts.
\n\n\n## Reference(s):
\n- [Fulcra Memory CLI Reference](references/fulcra-memory-cli.md)
\n- [ClawHub Skill Page](https://clawhub.ai/fulcra/fulcra-memory)
\n\n\n## Skill Output:
\n**Output Type(s):** [text, markdown, code, shell commands, configuration, guidance]
\n**Output Format:** [Markdown guidance with shell command examples and file path conventions]
\n**Output Parameters:** [1D]
\n**Other Properties Related to Output:** [Produces instructions for Fulcra CLI memory backup, restore, rollback, cloning, artifact upload, and team inbox workflows.]
\n\n## Skill Version(s):
\n0.0.1-beta (source: server release evidence)
\n\n## Ethical Considerations:
\nUsers should evaluate whether this skill is appropriate for their environment, review any generated or modified files before relying on them, and apply their organization's safety, security, and compliance requirements before deployment.
\n\nArchive v0.0.1: 4 files, 5669 bytes\n\nFiles: references/fulcra-memory-cli.md (4742b), skill-card.md (2541b), SKILL.md (4470b), _meta.json (132b)\n\nFile v0.0.1:SKILL.md\n\n---\nname: fulcra-memory\ndescription: \"Manages agent memory backup, restoration, rollback, and cloning using Fulcra's versioned file storage.\"\nhomepage: \"https://github.com/fulcradynamics/agent-skills\"\nlicense: \"MIT\"\nuser-invocable: true\nmetadata: { \"openclaw\": { \"emoji\": \"🧠\" } }\n---\n\n# Fulcra Agent Memory Management\n\nThe **primary role** of this skill is to enable an agent to securely back up its memory so it can be rolled back, cloned, or monitored (via \"top of mind\") using Fulcra's versioned file storage.\n\nBecause an agent's memory (e.g., `MEMORY.md`, `IDENTITY.md`, `SOUL.md`, and the `memory/` logs) evolves dynamically, periodically saving this state ensures that no context is lost and allows the user to safely rewind the agent if a task goes off track.\n\n## Primary Role: Memory & State Management\n\n### 1. The Memory Namespace\nFor core memory operations, agents use the standardized prefix: `agent//memory/`.\nThis dedicated directory tracks the agent's state across two main files: the `memory.gz` backup archive and a `top_of_mind.md` status file.\n\n### 2. Periodic Backups & Top of Mind\nAgents should run a backup process periodically (e.g., triggered by `HEARTBEAT.md`). Every time an agent backs up its state:\n1. It creates and uploads the compressed `memory.gz` archive.\n2. It writes a concise `top_of_mind.md` file locally—recording its current tasks, context, and what it's thinking about—and uploads it directly to `agent//memory/top_of_mind.md`. This gives the user immediate visibility into the agent's current state.\n\n### 3. Versioned Storage\nFulcra's file upload system inherently versions files uploaded to the same path. \n- The target path structure for backups is: `agent//memory/memory.gz`\n- By repeatedly uploading to this exact same path, Fulcra creates a historical timeline of the agent's memory states.\n\n### 4. Safe Rollbacks (The \"Undo\" Requirement)\nIf a user asks to roll back or restore memory from a previous date/version, **the agent MUST immediately upload a fresh backup of its current state BEFORE executing the restore.** This guarantees that if the user changes their mind, they can easily \"undo\" the rollback.\n\n### 5. Agent Cloning\nBy pointing the download command to a different agent's path (e.g., `agent//memory/memory.gz`), an agent can effectively clone another agent's memories and identity.\n\n## Advanced Extensions: Artifacts & Teamwork\n\n**Important Context:** The following features are powerful extensions of the agent's capabilities. However, they should only be presented or utilized when directly relevant to the user's current workflow. For example, if a user is simply going through initial onboarding via `fulcra-onboarding`, do not overwhelm them by bringing up team coordination unless they ask for it.\n\n### 6. Uploading User Artifacts\nAgents can store generated assets or dashboards created for the user.\n- **Namespace:** `agent//artifacts/` (e.g., `agent/wazir/artifacts/onboarding-dashboard.html`)\n- **Note:** Always ask for explicit user approval before uploading anything to the artifacts directory.\n\n### 7. Team Coordination & Shared Memory\nAgents can collaborate and share memory using a shared `team//` prefix in the Fulcra datastore.\nWithin a team's directory, the following structure is used:\n- **`team//artifacts/`**: Shared output files and deliverables created by the team.\n- **`team///inbox/`**: A drop-zone where other agents or users can place tasks, messages, or context for a specific agent.\n- **`team///archive/`**: Where an agent moves its inbox messages once they have been read and processed.\n\n**The Inbox Lifecycle:**\nWhen collaborating, agents write markdown messages to one another's inboxes. When the target agent processes its inbox, it must first upload the message to its `archive/` directory, and then delete the original file from its `inbox/`. Because Fulcra's file system is versioned, it automatically keeps a perfect audit trail of when the file was created in the inbox and when it was completed (deleted).\n\n## Workflow\n\nTo perform memory operations, agents must interact with the Fulcra CLI. \n\nSee the reference documentation for the exact commands needed to compress files, upload to Fulcra, and trigger restorations:\n- Read `references/fulcra-memory-cli.md` for exact file management and CLI execution steps.\n\nFile v0.0.1:_meta.json\n\n{\n \"ownerId\": \"kn731kt7vx4k5e3j56fpd4vep5874nrk\",\n \"slug\": \"fulcra-memory\",\n \"version\": \"0.0.1\",\n \"publishedAt\": 1780588951412\n}\n\nFile v0.0.1:references/fulcra-memory-cli.md\n\n---\nname: fulcra-memory-cli\ndescription: \"CLI command references for executing the memory backup, restore, and cloning operations with Fulcra.\"\n---\n\n# Fulcra Memory CLI Reference\n\nThis reference dictates the exact shell commands required to execute the `fulcra-memory` skill's operations. Ensure all tar and CLI operations run in the agent's root workspace (`~/.openclaw/workspace`).\n\n## 1. Creating a Backup and Uploading\n\nTo back up the agent's memory, you must generate a `top_of_mind.md` summary, compress the core identity files, and upload both to Fulcra.\n\n**Step A: Create Top of Mind**\nGenerate a concise markdown file summarizing your current tasks, thoughts, and context.\n```bash\n# Ensure you are in the workspace\ncd ~/.openclaw/workspace\nmkdir -p memory\necho \"# Current Agent State...\" > memory/top_of_mind.md\n# (Write your actual summary content to this file)\n```\n\n**Step B: Compress the files**\n```bash\n# Create a gzip tarball containing the essential memory files (ignore if some are missing)\ntar -czvf /tmp/memory.gz SOUL.md IDENTITY.md MEMORY.md memory/ 2>/dev/null || true\n```\n\n**Step C: Upload to Fulcra**\nUpload the files using the standardized agent path convention. Determine the agent's name (lowercase) to use in the path.\n\n```bash\n# Replace with the agent's actual name (e.g., treecle, wazir) in lowercase\nuv tool run fulcra-api file upload /tmp/memory.gz \"agent//memory/memory.gz\"\nuv tool run fulcra-api file upload memory/top_of_mind.md \"agent//memory/top_of_mind.md\"\n```\n\n## 2. Listing Memory History\n\nBecause Fulcra versions files automatically, you can see all previous backups of the memory using the `stat` command.\n\n```bash\nuv tool run fulcra-api file stat \"agent//memory/memory.gz\"\n```\n*(This command will output information about the file, including all previously uploaded versions and their UUIDs. Present these to the user so they can select a version to restore.)*\n\n## 3. Safe Restoration / Rollback\n\n**CRITICAL:** Before performing a restore, you MUST perform a fresh backup (Step 1 above) to ensure the current state isn't lost.\n\nOnce the pre-restore backup is complete, use the Fulcra CLI to set the active version, then download and extract it.\n\n**Step A: Restore the version in Fulcra**\n```bash\n# Instruct Fulcra to make the older version the active file\nuv tool run fulcra-api file restore \n```\n\n**Step B: Download the restored file**\n```bash\nuv tool run fulcra-api file download \"agent//memory/memory.gz\" /tmp/restored_memory.gz\n```\n\n**Step C: Extract and overwrite local memory**\n```bash\ncd ~/.openclaw/workspace\ntar -xzvf /tmp/restored_memory.gz\n```\n*(This will overwrite the local `SOUL.md`, `IDENTITY.md`, `MEMORY.md`, and the `memory/` directory with the state from the downloaded archive.)*\n\n## 4. Cloning Another Agent's Memory\n\nTo clone, skip the restore step and simply download the target agent's `memory.gz` archive.\n\n```bash\n# Download the target agent's memory\nuv tool run fulcra-api file download \"agent//memory/memory.gz\" /tmp/restored_memory.gz\n\n# Extract locally (Overwrites current identity and memory)\ncd ~/.openclaw/workspace\ntar -xzvf /tmp/restored_memory.gz\n```\n\n## 5. Uploading User Artifacts\n\nWhen an agent generates a file (like an HTML dashboard, an image, or a report), and the user explicitly approves saving it to their Fulcra account, upload it to the `artifacts/` subdirectory.\n\n```bash\n# Replace with the agent's name, and with the file's name\nuv tool run fulcra-api file upload /path/to/local/file \"agent//artifacts/\"\n```\n\n## 6. Team Coordination (Inbox & Archive)\n\nAgents can coordinate by writing to and reading from team namespaces.\n\n**Step A: Sending a message to a teammate's inbox**\n```bash\n# Upload a local markdown file to the target agent's inbox\nuv tool run fulcra-api file upload /tmp/task-for-wazir.md \"team///inbox/task_123.md\"\n```\n\n**Step B: Checking your inbox**\n```bash\n# List files in your agent's inbox\nuv tool run fulcra-api file list \"team///inbox/\"\n```\n\n**Step C: Processing and Archiving a message**\nOnce you have downloaded and read a message from your inbox, move it to the archive.\n\n```bash\n# 1. Download to read (if you haven't already)\nuv tool run fulcra-api file download \"team///inbox/task_123.md\" /tmp/task_123.md\n\n# 2. Upload it to your archive directory\nuv tool run fulcra-api file upload /tmp/task_123.md \"team///archive/task_123.md\"\n\n# 3. Delete it from the inbox to clear it\nuv tool run fulcra-api file delete \"team///inbox/task_123.md\"\n```\n\nFile v0.0.1:skill-card.md\n\n## Description:
\nManages agent memory backup, restoration, rollback, and cloning using Fulcra's versioned file storage.
\n\nThis skill is ready for commercial/non-commercial use.
\n\n## Publisher:
\n[fulcra](https://clawhub.ai/user/fulcra)
\n\n### License/Terms of Use:
\nMIT-0
\n\n\n## Use Case:
\nDevelopers and agent operators use this skill to save, inspect, restore, roll back, and clone agent memory through Fulcra's versioned file storage. It also supports approved artifact uploads and team inbox workflows when those are relevant to the user's work.
\n\n### Deployment Geography for Use:
\nGlobal
\n\n## Known Risks and Mitigations:
\nRisk: Memory restore, rollback, or cloning can overwrite an agent's local identity and memory state.
\nMitigation: Require explicit user approval and create a fresh backup before any restore, rollback, or clone operation.
\nRisk: Backups and artifact uploads can expose private user data or secrets if sensitive files are stored in memory paths.
\nMitigation: Use only trusted Fulcra accounts, avoid storing secrets or private user data in memory files, and review files before upload.
\nRisk: Downloaded or cloned archives can introduce untrusted memory or identity content into a live workspace.
\nMitigation: Inspect restored or cloned archives before extraction and only apply archives from agents and accounts the user is allowed to access.
\nRisk: Team inbox deletion can remove active task messages before they are safely retained.
\nMitigation: Archive each inbox message before deletion and require user approval for destructive inbox changes.
\n\n\n## Reference(s):
\n- [Fulcra Memory CLI Reference](references/fulcra-memory-cli.md)
\n\n\n## Skill Output:
\n**Output Type(s):** [Markdown, Shell commands, Configuration guidance, Files]
\n**Output Format:** [Markdown guidance with inline shell commands and file path conventions]
\n**Output Parameters:** [1D]
\n**Other Properties Related to Output:** [Produces and manages compressed memory archives, top-of-mind summaries, restored files, uploaded artifacts, and team inbox messages through Fulcra CLI workflows.]
\n\n## Skill Version(s):
\n0.0.1 (source: server release evidence)
\n\n## Ethical Considerations:
\nUsers should evaluate whether this skill is appropriate for their environment, review any generated or modified files before relying on them, and apply their organization's safety, security, and compliance requirements before deployment.
","readmeExcerpt":"Skill: fulcra-memory Owner: fulcra Summary: Manages agent progress reporting and OKF-compliant memory syncing to Fulcra. Tags: latest:0.1.9 Version history: v0.1.9 | 2026-08-12T22:24:30.127Z | auto - Updated reference links in documentation to point to fulcra-get-started instead of fulcra-onboarding. - Removed the unused skill-card.md file. v0.0.8 | 2026-07-01T22:05:28.915Z | auto **Major update: Skill refocused on p","codeSnippets":[],"executableExamples":[{"language":"bash","snippet":"# Get a summary of files changed in the last 1 day\nuv tool run fulcra-api data-updates \"1 day\"\n\n# Example output:\n# {\n# \"data_types\": {},\n# \"file_changes\": [\n# {\n# \"full_name\": \"/agent/treecle/memory/knowledge/programming/python.md\",\n# \"uploaded_at\": \"2026-07-01T21:23:28.690719Z\",\n# \"state\": \"uploaded\",\n# \"...\": \"...\"\n# },\n# {\n# \"full_name\": \"/agent/treecle/memory/session/20260701-120000_setup.md\",\n# \"uploaded_at\": \"2026-07-01T21:23:28.690719Z\",\n# \"state\": \"uploaded\",\n# \"...\": \"...\"\n# }\n# ]\n# }"},{"language":"bash","snippet":"# Ensure you are in the workspace\ncd ~/.openclaw/workspace\nmkdir -p memory\n\n# 1. Create progress.md\ncat << 'EOF' > memory/progress.md\n---\ntype: Progress Report\ntitle: Agent Progress\n---\n# Recent Accomplishments\n...\nEOF\n\n# 2. Update log.md\nDATE=$(date -u +\"%Y-%m-%d\")\necho \"## $DATE\" >> memory/log.md\necho \"* **Sync**: Updated progress report.\" >> memory/log.md\n\n# 3. Create index.md if it doesn't exist\ncat << 'EOF' > memory/index.md\n# Memory Namespace\n\n* [Progress Report](./progress.md) - The latest summary of agent activities.\n* [Update Log](./log.md) - History of changes to this namespace.\nEOF"},{"language":"bash","snippet":"# Replace with the agent's actual name (e.g., treecle, wazir) in lowercase\nuv tool run fulcra-api file upload memory/progress.md \"agent//progress.md\"\nuv tool run fulcra-api file upload memory/log.md \"agent//log.md\"\nuv tool run fulcra-api file upload memory/index.md \"agent//index.md\""},{"language":"bash","snippet":"uv tool run fulcra-api file upload memory/role.md \"agent//role.md\""},{"language":"bash","snippet":"uv tool run fulcra-api file upload memory/session/20260623-180530_setup-dashboard.md \"agent//session/20260623-180530_setup-dashboard.md\""},{"language":"text","snippet":"## 4. Personal Inbox Lifecycle\n\nAgents can receive files in their personal inbox (`inbox/`) from users or external triggers. Users can drop files here manually without needing strict timestamp formats.\n\n**Step A: Process and Archive**\nRead the file from the inbox. Once processed, you must upload it to the `archive/` directory. If the original file name doesn't start with a timestamp, you **must prepend one** (`YYYYMMDD-HHMMSS_`)."}],"parameters":null,"dependencies":[],"permissions":[],"extractedFiles":[{"path":"SKILL.md","content":"---\nname: fulcra-memory\ndescription: \"Manages agent progress reporting and OKF-compliant memory syncing to Fulcra.\"\nhomepage: \"https://github.com/fulcradynamics/agent-skills\"\nlicense: \"MIT\"\nuser-invocable: true\nmetadata: { \"openclaw\": { \"emoji\": \"🧠\" } }\n---\n\n# Fulcra Agent Memory Sync\n\nThe **primary role** of this skill is to help the agent keep track of what it is doing in a readable and transferable way using Fulcra's versioned file storage, strictly following the Open Knowledge Format (OKF).\n\nBy periodically logging activities to `progress.md` and keeping the memory namespace structured, other agents and users can easily see what this agent has accomplished and what it is working on next.\n\n*(Note: For agent backup, rollback, and cloning, see the `fulcradynamics/agent-skills/fulcra-agent-backup` skill.)*\n\n## Primary Role: Progress & OKF Compliance\n\n### 1. The Memory Namespace (OKF Compliant)\nFor core memory tracking, agents use the standardized prefix: `agent//`.\nThis dedicated directory tracks the agent's state. It must conform to the Open Knowledge Format (OKF), meaning it contains an `index.md`, a `log.md`, and markdown concept files like `progress.md`.\n\n**IMPORTANT OKF EFFICIENCY DIRECTIVE:** While OKF compliance is mandatory, it must not become cumbersome.\n- **DO NOT** attempt to index every individual file, especially not transient or highly repetitive files like individual messages inside an `inbox/` subdirectory. \n- For directories that contain many small or transient items (like an `inbox/` or `queue/`), simply list the directory itself in the `index.md` with a high-level description of its purpose (e.g., \"Contains unread messages from the web dashboard\").\n- Keep the `index.md` and `log.md` focused on top-level structure and significant milestones (like the `progress.md`) so they remain useful but lightweight.\n\n### 2. Discovering Recent Memory Changes\nRather than exhaustively querying your namespace to remember what you or others recently learned, use the Fulcra API's `data-updates` command (e.g., `uv tool run fulcra-api data-updates \"1 day\"`). This will return a summary of all files that changed recently, allowing you to easily see what knowledge was added and which memory files were updated across your local and team-shared namespaces.\n\n### 3. Progress Reports\nAgents should run a progress sync periodically (e.g., triggered by `HEARTBEAT.md`). Every time an agent syncs its state:\n1. It writes a concise `progress.md` report locally—summarizing the work it has recently done, and what it is planning to do next—and uploads it directly to `agent//progress.md`. The `progress.md` file must include proper OKF YAML frontmatter (with `type: Progress Report`). This file is a report specifically designed to let the user review the agent's recent accomplishments and intended next steps. It is **not** intended to dump internal agent state, chain-of-thought, or precise working memory.\n2. It updates the `log.md` (Up"},{"path":"_meta.json","content":"{\n \"ownerId\": \"kn731kt7vx4k5e3j56fpd4vep5874nrk\",\n \"slug\": \"fulcra-memory\",\n \"version\": \"0.1.9\",\n \"publishedAt\": 1786573470127\n}"},{"path":"references/fulcra-memory-cli.md","content":"---\nname: fulcra-memory-cli\ndescription: \"CLI command references for executing memory sync and progress reporting with Fulcra.\"\n---\n\n# Fulcra Memory CLI Reference\n\nThis reference dictates the exact shell commands required to execute the `fulcra-memory` skill's operations. Ensure all CLI operations run in the agent's root workspace (`~/.openclaw/workspace`).\n\n## Authentication Note\nIf you need to authenticate to Fulcra before running these commands, you must use the non-blocking two-step login process to prevent the CLI from hanging:\n1. `uv tool run fulcra-api auth login --get-auth-url` (present URL and code to user)\n2. `uv tool run fulcra-api auth login --device-code --poll-timeout=5` (after user finishes flow)\n\n## 1. Discovering Recent Memory Changes\n\nTo quickly see what memory files were updated or new knowledge was added recently:\n\n```bash\n# Get a summary of files changed in the last 1 day\nuv tool run fulcra-api data-updates \"1 day\"\n\n# Example output:\n# {\n# \"data_types\": {},\n# \"file_changes\": [\n# {\n# \"full_name\": \"/agent/treecle/memory/knowledge/programming/python.md\",\n# \"uploaded_at\": \"2026-07-01T21:23:28.690719Z\",\n# \"state\": \"uploaded\",\n# \"...\": \"...\"\n# },\n# {\n# \"full_name\": \"/agent/treecle/memory/session/20260701-120000_setup.md\",\n# \"uploaded_at\": \"2026-07-01T21:23:28.690719Z\",\n# \"state\": \"uploaded\",\n# \"...\": \"...\"\n# }\n# ]\n# }\n```\n\n*Note: The `file_changes` key is a list of file metadata objects. You can extract the `full_name` from each to see the file paths. If the summary shows that specific memory files were changed, you can then read those specific files to update your context.*\n\n## 2. Syncing Progress and OKF Files\n\nTo keep the agent's memory in sync, generate a `progress.md` summary, ensure OKF files are updated, and upload them to Fulcra.\n\n**Step A: Create Progress Report and OKF Files**\nGenerate a concise markdown file summarizing the work you have recently completed and what you plan to do next. It must include OKF YAML frontmatter. Do not include internal state or chain-of-thought. Also ensure `index.md` and `log.md` are created or updated.\n\n```bash\n# Ensure you are in the workspace\ncd ~/.openclaw/workspace\nmkdir -p memory\n\n# 1. Create progress.md\ncat << 'EOF' > memory/progress.md\n---\ntype: Progress Report\ntitle: Agent Progress\n---\n# Recent Accomplishments\n...\nEOF\n\n# 2. Update log.md\nDATE=$(date -u +\"%Y-%m-%d\")\necho \"## $DATE\" >> memory/log.md\necho \"* **Sync**: Updated progress report.\" >> memory/log.md\n\n# 3. Create index.md if it doesn't exist\ncat << 'EOF' > memory/index.md\n# Memory Namespace\n\n* [Progress Report](./progress.md) - The latest summary of agent activities.\n* [Update Log](./log.md) - History of changes to this namespace.\nEOF\n```\n\n**Step B: Upload to Fulcra**\nUpload the files using the standardized agent path convention. Determine the agent's name (lowercase) to use in the path.\n\n```bash\n# Replace with the agent's actual n"},{"path":"skill-card.md","content":"## Description:\n\nManages agent progress reporting and OKF-compliant memory syncing to Fulcra.\n\nThis skill is ready for commercial/non-commercial use.\n\n## Publisher:\n\n[fulcra](https://clawhub.ai/user/fulcra)\n\n### License/Terms of Use:\n\nMIT-0\n\n## Use Case:\n\nDevelopers and agents use this skill to maintain OKF-compliant progress, role, session, task, knowledge, and inbox memory files and sync selected summaries to Fulcra.\n\n### Deployment Geography for Use:\n\nGlobal\n\n## Known Risks and Mitigations:\n\nRisk: The skill uploads persistent agent and user context, including work summaries, decisions, links, preferences, and task context, to Fulcra.\n\nMitigation: Review memory content before upload, minimize sensitive disclosures, and require explicit user approval before syncing sensitive context.\n\nRisk: The skill relies on an external Fulcra CLI for authentication, uploads, and deletion, and the security evidence notes that the CLI is unpinned.\n\nMitigation: Use a pinned and trusted fulcra-api installation and limit Fulcra credentials to the intended agent namespace.\n\nRisk: Inbox processing can delete source files after archiving.\n\nMitigation: Verify archive upload success before deleting inbox files.\n\n## Reference(s):\n\n- [Fulcra Memory CLI Reference](references/fulcra-memory-cli.md)\n- [Fulcra CLI Documentation](https://raw.githubusercontent.com/fulcradynamics/agent-skills/main/skills/fulcra-get-started/references/fulcra-cli.md)\n- [Open Knowledge Format Specification](https://github.com/GoogleCloudPlatform/knowledge-catalog/blob/main/okf/SPEC.md)\n- [ClawHub Skill Page](https://clawhub.ai/fulcra/skills/fulcra-memory)\n\n## Skill Output:\n\n**Output Type(s):** [text, markdown, shell commands, configuration, guidance]\n\n**Output Format:** [Markdown guidance with inline shell commands and OKF-formatted memory file examples]\n\n**Output Parameters:** [1D]\n\n**Other Properties Related to Output:** [May result in persistent Fulcra memory files when the agent follows the CLI workflow.]\n\n## Skill Version(s):\n\n0.1.9 (source: release metadata)\n\n## Ethical Considerations:\n\nUsers should evaluate whether this skill is appropriate for their environment, review any generated or modified files before relying on them, and apply their organization's safety, security, and compliance requirements before deployment."}],"languages":[],"docsSourceLabel":"CLAWHUB","editorialOverview":null,"editorialQuality":{"score":100,"threshold":65,"status":"thin","wordCount":1917,"uniquenessScore":40,"reasons":["uniqueness-below-45"]}},"media":{"evidence":{"source":"no-media","verified":false,"confidence":"low","updatedAt":"2026-10-11T12:06:09.927Z","emptyReason":"No screenshots, media assets, or demo links are available."},"primaryImageUrl":null,"mediaAssetCount":0,"assets":[],"demoUrl":null},"ownerResources":{"evidence":{"source":"unclaimed","verified":false,"confidence":"low","updatedAt":"2026-10-11T12:06:09.927Z","emptyReason":"This page has not been claimed by the agent owner."},"hasCustomPage":false,"customPageUpdatedAt":null,"customLinks":[],"structuredLinks":{"docsUrl":null,"demoUrl":null,"supportUrl":null,"pricingUrl":null,"statusUrl":null},"customPage":null},"relatedAgents":{"evidence":{"source":"protocol-neighbors","verified":false,"confidence":"medium","updatedAt":"2026-10-11T15:15:28.910Z","emptyReason":null},"items":[{"id":"8ebccd8e-3863-4187-8355-c3f14e1f9edf","entityType":"agent","canonicalPath":"/agent/iofficeai-aionui","slug":"iofficeai-aionui","name":"AionUi","description":"Free, local, open-source 24/7 Cowork app and OpenClaw for Gemini CLI, Claude Code, Codex, OpenCode, Qwen Code, Goose CLI, Auggie, and more | 🌟 Star if you like it!","url":"https://github.com/iOfficeAI/AionUi","homepage":"https://www.aionui.com","source":"GITHUB_REPOS","protocols":["MCP","OPENCLAW"],"capabilities":[],"safetyScore":100,"overallRank":70,"updatedAt":"2026-10-09T19:11:12.944Z","createdAt":"2026-02-25T03:38:16.584Z","downloads":null},{"id":"b917f68a-ebff-438e-84f8-3f4b2494c0bc","entityType":"agent","canonicalPath":"/agent/activepieces-activepieces","slug":"activepieces-activepieces","name":"activepieces","description":"AI Agents & MCPs & AI Workflow Automation • (~400 MCP servers for AI agents) • AI Automation / AI Agent with MCPs • AI Workflows & AI Agents • MCPs for AI Agents","url":"https://github.com/activepieces/activepieces","homepage":"https://www.activepieces.com","source":"GITHUB_REPOS","protocols":["OPENCLAW"],"capabilities":[],"safetyScore":100,"overallRank":70,"updatedAt":"2026-04-15T02:22:12.426Z","createdAt":"2026-02-25T03:38:12.412Z","downloads":null},{"id":"5cb26759-3a39-483f-94cf-276a98c13bb8","entityType":"agent","canonicalPath":"/agent/cherryhq-cherry-studio","slug":"cherryhq-cherry-studio","name":"cherry-studio","description":"AI productivity studio with smart chat, autonomous agents, and 300+ assistants. Unified access to frontier LLMs","url":"https://github.com/CherryHQ/cherry-studio","homepage":"https://cherry-ai.com","source":"GITHUB_REPOS","protocols":["MCP","OPENCLAW"],"capabilities":[],"safetyScore":100,"overallRank":70,"updatedAt":"2026-04-11T14:38:40.986Z","createdAt":"2026-02-25T03:38:19.379Z","downloads":null},{"id":"6f6582d0-5d76-4f0f-b81d-86520247950b","entityType":"agent","canonicalPath":"/agent/copilotkit-copilotkit","slug":"copilotkit-copilotkit","name":"CopilotKit","description":"The Frontend for Agents & Generative UI. React + Angular","url":"https://github.com/CopilotKit/CopilotKit","homepage":"https://docs.copilotkit.ai","source":"GITHUB_REPOS","protocols":["OPENCLAW"],"capabilities":[],"safetyScore":100,"overallRank":70,"updatedAt":"2026-03-25T09:50:57.846Z","createdAt":"2026-02-25T03:39:14.617Z","downloads":null}],"links":{"hub":"/agent","source":"/agent/source/clawhub","protocols":[{"label":"OpenClaw","href":"/agent/protocol/openclew"}]}}}