{"id":"c54a0139-1f27-470d-9e70-379306ad7607","entityType":"agent","slug":"clawhub-adelpro-continue-learning","name":"continue-learning","canonicalUrl":"https://www.xpersona.co/agent/clawhub-adelpro-continue-learning","canonicalPath":"/agent/clawhub-adelpro-continue-learning","generatedAt":"2026-10-10T11:49:38.740Z","source":"CLAWHUB","claimStatus":"UNCLAIMED","verificationTier":"NONE","summary":{"evidence":{"source":"CLAWHUB","verified":false,"confidence":"medium","updatedAt":"2026-10-10T09:10:22.804Z","emptyReason":null},"description":"Instinct-based learning system for OpenClaw. Analyzes sessions, detects patterns, creates atomic learnings with confidence scoring, and suggests optimizations for self-evolution. Works alongside agent-self-improvement for complete learning: internal session analysis + external user feedback. Use when: you want your AI agent to learn from its own behavior, improve over time, discover optimization opportunities, or build a self-improving automation system. Don't use when: static agent behavior is preferred.","descriptionLabel":"Source description","evidenceSummary":"Capability contract not published. No trust telemetry is available yet. 1.5K downloads reported by the source. Last updated 10/10/2026.","installCommand":"clawhub skill install s177vajy8ffpjnym8evc1jcxr5877x59:continue-learning","sourceUrl":"https://clawhub.ai/adelpro/continue-learning","homepage":"https://clawhub.ai/adelpro/skills/continue-learning","primaryLinks":[{"label":"View on ClawHub","url":"https://clawhub.ai/adelpro/continue-learning","kind":"source"},{"label":"Homepage","url":"https://clawhub.ai/adelpro/skills/continue-learning","kind":"homepage"}],"safetyScore":84,"overallRank":62,"popularityScore":64,"trustScore":null,"claimedByName":null,"isOwner":false,"seoDescription":"continue-learning technical dossier on Xpersona with agent coverage, OPENCLEW support, and live trust metadata."},"coverage":{"evidence":{"source":"public-profile","verified":false,"confidence":"medium","updatedAt":"2026-10-10T09:10:22.804Z","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":"2026-10-10T09:10:22.804Z","emptyReason":null},"stars":null,"forks":null,"downloads":1534,"packageName":null,"latestVersion":"1.3.2","tractionLabel":"1.5K downloads"},"release":{"evidence":{"source":"CLAWHUB","verified":false,"confidence":"medium","updatedAt":"2026-10-10T09:10:22.804Z","emptyReason":null},"lastUpdatedAt":"2026-10-10T09:10:22.804Z","lastCrawledAt":"2026-10-10T09:10:22.804Z","lastIndexedAt":null,"nextCrawlAt":"2026-10-11T09:10:22.804Z","lastVerifiedAt":null,"highlights":[{"version":"1.3.2","createdAt":"2026-08-20T19:41:13.454Z","changelog":"Version 1.3.2 of continue-learning adds security and privacy controls for learning data. - Session analysis now requires explicit agent scoping (`--agent <name>` or `--all`); running with no scope is not allowed. - All session-derived data stored is redacted for secrets and Unicode/control stripped before saving. - Retention limits added: instincts, evidence, and patterns are capped to recent data; nothing is retained unbounded. - New `prune` command deletes all stored learning data (instincts, patterns, optimizations). - Raw transcripts are never persisted; only redacted and aggregated information is stored. - skill-card.md documentation file removed.","fileCount":5,"zipByteSize":10515},{"version":"1.3.1","createdAt":"2026-08-20T19:22:46.898Z","changelog":"- Renamed the skill from \"openclaw-continuous-learning\" to \"continue-learning\". - Added a new README.md file. - Removed the skill-card.md file. - Updated SKILL.md with the new skill name and version.","fileCount":5,"zipByteSize":8725},{"version":"1.3.0","createdAt":"2026-03-14T17:50:18.226Z","changelog":"- Updated documentation to clarify session source directory (`~/.openclaw/agents/`) and improved quick start instructions. - Enhanced example commands and added a command for showing error patterns. - Minor version bump and metadata update for improved usability and clearer guidance.","fileCount":4,"zipByteSize":8571},{"version":"1.2.0","createdAt":"2026-03-14T09:00:29.051Z","changelog":"Version 1.2.0 - Added explicit integration guidance for use alongside agent-self-improvement, enabling combined internal (session analysis) and external (user feedback) learning. - Updated documentation to clarify skill boundaries and workflow for collaborative learning. - No changes to APIs or storage—improvement is purely in documentation and usability clarity.","fileCount":3,"zipByteSize":5799},{"version":"1.1.0","createdAt":"2026-02-25T09:17:03.763Z","changelog":"openclaw-continuous-learning 1.1.0 - Enhanced description and documentation, detailing instinct-based learning, pattern detection, and confidence scoring. - Clarified use cases, including agent self-improvement, user preference learning, performance optimization, and error pattern detection. - Added comprehensive setup and architecture sections with examples and best practices. - Provided explanation of storage structure and confidence scoring system. - Included FAQ and related skills for better onboarding and context.","fileCount":3,"zipByteSize":5449}]},"execution":{"evidence":{"source":"CLAWHUB","verified":false,"confidence":"low","updatedAt":null,"emptyReason":"No published capability contract is available yet."},"installCommand":"clawhub skill install s177vajy8ffpjnym8evc1jcxr5877x59:continue-learning","setupComplexity":"low","setupSteps":["Install using `clawhub skill install s177vajy8ffpjnym8evc1jcxr5877x59:continue-learning` 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/adelpro/continue-learning 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-adelpro-continue-learning/snapshot","contractUrl":"https://www.xpersona.co/api/v1/agents/clawhub-adelpro-continue-learning/contract","trustUrl":"https://www.xpersona.co/api/v1/agents/clawhub-adelpro-continue-learning/trust"},"curlExamples":["curl -s \"https://www.xpersona.co/api/v1/agents/clawhub-adelpro-continue-learning/snapshot\"","curl -s \"https://www.xpersona.co/api/v1/agents/clawhub-adelpro-continue-learning/contract\"","curl -s \"https://www.xpersona.co/api/v1/agents/clawhub-adelpro-continue-learning/trust\""],"jsonRequestTemplate":{"query":"summarize this repo","constraints":{"maxLatencyMs":2000,"protocolPreference":["OPENCLEW"]}},"jsonResponseTemplate":{"ok":true,"result":{"summary":"...","confidence":0.9},"meta":{"source":"CLAWHUB","generatedAt":"2026-10-10T11:49:38.738Z"}},"retryPolicy":{"maxAttempts":3,"backoffMs":[500,1500,3500],"retryableConditions":["HTTP_429","HTTP_503","NETWORK_TIMEOUT"]}},"endpoints":{"dossierUrl":"https://www.xpersona.co/api/v1/agents/clawhub-adelpro-continue-learning/dossier","snapshotUrl":"https://www.xpersona.co/api/v1/agents/clawhub-adelpro-continue-learning/snapshot","contractUrl":"https://www.xpersona.co/api/v1/agents/clawhub-adelpro-continue-learning/contract","trustUrl":"https://www.xpersona.co/api/v1/agents/clawhub-adelpro-continue-learning/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":"2026-10-10T09:10:22.804Z","emptyReason":null},"readme":"Skill: continue-learning\n\nOwner: adelpro\n\nSummary: Instinct-based learning system for OpenClaw. Analyzes sessions, detects patterns, creates atomic learnings with confidence scoring, and suggests optimizations for self-evolution. Works alongside agent-self-improvement for complete learning: internal session analysis + external user feedback. Use when: you want your AI agent to learn from its own behavior, improve over time, discover optimization opportunities, or build a self-improving automation system. Don't use when: static agent behavior is preferred.\n\nTags: latest:1.3.2\n\nVersion history:\n\nv1.3.2 | 2026-08-20T19:41:13.454Z | auto\n\nVersion 1.3.2 of continue-learning adds security and privacy controls for learning data.\n\n- Session analysis now requires explicit agent scoping (`--agent <name>` or `--all`); running with no scope is not allowed.\n- All session-derived data stored is redacted for secrets and Unicode/control stripped before saving.\n- Retention limits added: instincts, evidence, and patterns are capped to recent data; nothing is retained unbounded.\n- New `prune` command deletes all stored learning data (instincts, patterns, optimizations).\n- Raw transcripts are never persisted; only redacted and aggregated information is stored.\n- skill-card.md documentation file removed.\n\nv1.3.1 | 2026-08-20T19:22:46.898Z | auto\n\n- Renamed the skill from \"openclaw-continuous-learning\" to \"continue-learning\".\n- Added a new README.md file.\n- Removed the skill-card.md file.\n- Updated SKILL.md with the new skill name and version.\n\nv1.3.0 | 2026-03-14T17:50:18.226Z | auto\n\n- Updated documentation to clarify session source directory (`~/.openclaw/agents/`) and improved quick start instructions.\n- Enhanced example commands and added a command for showing error patterns.\n- Minor version bump and metadata update for improved usability and clearer guidance.\n\nv1.2.0 | 2026-03-14T09:00:29.051Z | auto\n\nVersion 1.2.0\n\n- Added explicit integration guidance for use alongside agent-self-improvement, enabling combined internal (session analysis) and external (user feedback) learning.\n- Updated documentation to clarify skill boundaries and workflow for collaborative learning.\n- No changes to APIs or storage—improvement is purely in documentation and usability clarity.\n\nv1.1.0 | 2026-02-25T09:17:03.763Z | auto\n\nopenclaw-continuous-learning 1.1.0\n\n- Enhanced description and documentation, detailing instinct-based learning, pattern detection, and confidence scoring.\n- Clarified use cases, including agent self-improvement, user preference learning, performance optimization, and error pattern detection.\n- Added comprehensive setup and architecture sections with examples and best practices.\n- Provided explanation of storage structure and confidence scoring system.\n- Included FAQ and related skills for better onboarding and context.\n\nArchive index:\n\nArchive v1.3.2: 5 files, 10515 bytes\n\nFiles: README.md (237b), scripts/analyze.mjs (15849b), skill-card.md (2121b), SKILL.md (8967b), _meta.json (136b)\n\nFile v1.3.2:SKILL.md\n\n---\nname: continue-learning\nslug: continue-learning\nversion: 1.3.2\ndescription: |\n  Instinct-based learning system for OpenClaw. Analyzes sessions, detects patterns, creates atomic learnings with confidence scoring, and suggests optimizations for self-evolution.\n  \n  Works alongside agent-self-improvement for complete learning: internal session analysis + external user feedback.\n  \n  Use when: you want your AI agent to learn from its own behavior, improve over time, discover optimization opportunities, or build a self-improving automation system.\n  \n  Don't use when: static agent behavior is preferred.\ntriggers:\n  - continuous learning\n  - self improving agent\n  - agent evolution\n  - pattern detection\n  - session analysis\n  - ai learning\n  - agent optimization\n  - automation improvement\n  - self evolution\nmetadata:\n  openclaw:\n    emoji: \"🧠\"\n    requires:\n      bins: [\"node\"]\n---\n\n# Continuous Learning for AI Agents\n\nAn instinct-based learning system that helps AI agents improve themselves through observation and pattern detection.\n\n## What This Skill Does\n\n- **Analyzes session history** - Reviews agent interactions and outputs\n- **Detects patterns** - Identifies recurring behaviors, preferences, workflows\n- **Creates instincts** - Atomic learnings with confidence scores\n- **Suggests optimizations** - Based on observed behavior patterns\n- **Enables self-evolution** - Converts insights into improvements\n\n## When to Use\n\n**Use when:**\n- Building self-improving AI agents\n- Want agent to learn from interactions\n- Discovering optimization opportunities\n- Creating adaptive automation\n- Tracking behavioral patterns\n\n**Skip when:**\n- Static, unchanging behavior preferred\n- No session history available\n- Simple, deterministic workflows only\n\n## Architecture\n\n```\n~/.openclaw/agents/ (session .jsonl files)\n        │\n        ▼\n┌───────────────────────────────────────────┐\n│ analyze.mjs                                │\n│ • Reads session history                   │\n│ • Extracts tool calls & errors             │\n│ • Detects patterns                         │\n└───────────────────────────────────────────┘\n        │\n        ▼\n┌───────────────────────────────────────────┐\n│ memory/learning/                           │\n│ • instincts.jsonl (atomic learnings)       │\n│ • patterns.json (aggregated)              │\n│ • optimizations.json (suggestions)         │\n└───────────────────────────────────────────┘\n```\n\n## External Feedback (Sub-Skill)\n\nThis skill works with **agent-self-improvement** (ClawHub) for external user feedback capture:\n\n- **Internal Learning**: Session analysis (this skill)\n- **External Learning**: User feedback via `SKILL:agent-self-improvement`\n\n### Combined Usage\n\n```\n# Nightly: Internal analysis\nSKILL:openclaw-continuous-learning --analyze\n\n# After any output: Capture feedback\nSKILL:agent-self-improvement --job <task> --feedback \"<user response>\"\n\n# Daily: Generate combined improvements\nSKILL:agent-self-improvement --improve all\n```\n\n### Feedback Flow\n\n```\nUser Response → agent-self-improvement → Directive Hints\n        ↓\nSession Analysis → openclaw-continuous-learning → Internal Patterns\n        ↓\nCombined Insights → Agent Optimization\n```\n\nBoth skills store learnings in `memory/learning/` and can reference each other's data.\n\n## Confidence Scoring\n\n| Score | Meaning | Behavior |\n|-------|---------|----------|\n| 0.3 | Tentative | Suggested but not enforced |\n| 0.5 | Moderate | Applied when relevant |\n| 0.7 | Strong | Auto-approved |\n| 0.9 | Core behavior | Always apply |\n\n**Confidence increases when:**\n- Pattern observed repeatedly\n- User doesn't correct behavior\n- Multiple observations agree\n\n**Confidence decreases when:**\n- User explicitly corrects\n- Pattern not observed recently\n- Contradicting evidence appears\n\n## Key Concepts\n\n### Instincts\n\nAn instinct is a small learned behavior:\n\n```yaml\nid: prefer-simplicity\ntrigger: \"when solving problems\"\nconfidence: 0.7\ndomain: problem_solving\n---\n# Prefer Simple Solutions\n\n## Action\nAlways choose the simplest solution that meets requirements.\n\n## Evidence\n- Observed preference for minimal code\n- User corrected over-engineered approaches\n```\n\n### Patterns\n\nAggregated observations grouped by category:\n- code_style\n- testing\n- git\n- debugging\n- workflow\n- communication\n\n### Optimizations\n\nActionable improvements derived from patterns.\n\n## Use Cases\n\n### 1. Agent Self-Improvement\n\n```\nAgent observes its own sessions:\n- What works consistently?\n- What gets corrected?\n- What patterns emerge?\n\nCreates instincts → Applies high-confidence patterns\n```\n\n### 2. User Preference Learning\n\n```\nLearn user preferences from interactions:\n- Coding style preferences\n- Communication preferences\n- Workflow preferences\n\nAdapt behavior accordingly\n```\n\n### 3. Performance Optimization\n\n```\nDetect performance patterns:\n- Slow operations\n- Bottlenecks\n- Optimization opportunities\n\nSuggest improvements\n```\n\n### 4. Error Pattern Detection\n\n```\nTrack error patterns:\n- Common failures\n- Resolution strategies\n- Prevention approaches\n\nBuild error-handling instincts\n```\n\n## Security & Privacy\n\n1. **Scoped analysis (opt-in).** `analyze.mjs` does not silently scan every agent. It analyses\n   only sessions of an agent you explicitly name with `--agent <name>`, or all agents only when\n   you pass `--all`. Running it with no scope refuses.\n2. **Redaction.** Any session-derived text stored (error snippets, evidence) is secret-redacted\n   and control/Unicode-stripped before persistence. Tokens, API keys, private-key blocks, long\n   hex, and URI credentials become `[REDACTED]`.\n3. **Retention caps.** Learning is bounded: instincts capped at the most recent 200, evidence\n   capped at 10 per instinct, patterns capped at the last 50 analysis runs. Nothing is retained\n   unbounded.\n4. **Deletion control.** Run `node scripts/analyze.mjs prune` to delete all stored instincts,\n   patterns, and optimizations at any time.\n5. **No raw transcripts.** Full session text is never persisted. Only aggregated counts and\n   redacted snippets are stored. Suggestions require your approval to apply — nothing\n   auto-modifies your agent.\n\n## Quick Start\n\n```bash\n# Analyze sessions for ONE explicit agent (opt-in scoping)\ncd ~/.openclaw/workspace/skills/continue-learning\nnode scripts/analyze.mjs --agent <agent-name>\n\n# Explicitly analyze all agents\nnode scripts/analyze.mjs --all\n\n# List learned instincts / optimizations / patterns (no scope needed)\nnode scripts/analyze.mjs instincts\nnode scripts/analyze.mjs list\nnode scripts/analyze.mjs patterns\n\n# Show redacted error patterns (scoped)\nnode scripts/analyze.mjs errors --agent <agent-name>\n\n# Delete all stored learning data\nnode scripts/analyze.mjs prune\n```\n\n## Setup\n\n### 1. Create storage directory\n\n```bash\nmkdir -p ~/.openclaw/workspace/memory/learning\n```\n\n### 2. Schedule analysis\n\nAdd to cron for periodic analysis:\n\n```json\n{\n  \"id\": \"continuous-learning\",\n  \"schedule\": \"0 22 * * *\"\n}\n```\n\n### 3. Integrate with daily tips\n\nConnect to daily summary for optimization delivery.\n\n## File Structure\n\n```\n~/.openclaw/workspace/\n└── memory/\n    └── learning/\n        ├── instincts.jsonl    # Atomic learnings\n        ├── patterns.json      # Aggregated patterns\n        └── optimizations.json # Suggestions\n```\n\n## Example Output\n\n```\n🧠 Learning Report\n\nPatterns Detected:\n- prefer-simplicity (0.7) ↑2\n- test-first (0.5) ↑1\n- commit-often (0.3) new\n\nConfidence Changes:\n- minimal-code: 0.5 → 0.7\n\nSuggested:\n1. Prioritize simple solutions\n2. Add pre-commit hooks\n3. Enable stricter typing\n```\n\n## Best Practices\n\n1. **Start simple** - Few patterns, low confidence\n2. **Validate often** - Check if patterns still hold\n3. **Review suggestions** - Don't auto-apply everything\n4. **Track confidence** - Update based on results\n5. **Export/share** - Build library of common patterns\n\n## FAQ\n\n**How is this different from memory?**\nMemory stores facts. This learns behavioral patterns and preferences.\n\n**How long to see results?**\nDepends on session volume. Typically 1-2 weeks for meaningful patterns.\n\n**Is it safe to auto-apply?**\nOnly high-confidence (0.7+) patterns. Always review suggestions first.\n\n## Related Skills\n\n- **skill-engineer** - Quality-gated skill development\n- **compound-engineering** - Session review and learning\n- **memory-setup** - Memory configuration\n- **openclaw-daily-tips** - Daily optimization tips\n\n---\n\n**Version:** 1.1.0  \n**Inspired by:** Anthropic's continuous learning patterns, Claude Code homunculus\n\nFile v1.3.2:README.md\n\n# continue-learning\n\nOpenClaw instinct-based learning system: session analysis, pattern detection, atomic learnings with confidence scoring, and optimization suggestions.\n\n## Install\n\n```bash\nnpx skills add adelpro/continue-learning\n```\n\nFile v1.3.2:_meta.json\n\n{\n  \"ownerId\": \"kn72cbvk9f5n48msm4t4sj3wyh80n2eb\",\n  \"slug\": \"continue-learning\",\n  \"version\": \"1.3.2\",\n  \"publishedAt\": 1787254873454\n}\n\nFile v1.3.2:skill-card.md\n\n## Description:\n\nInstinct-based learning system for OpenClaw that analyzes sessions, detects patterns, creates atomic learnings with confidence scoring, and suggests optimizations for self-evolution.\n\nThis skill is ready for commercial/non-commercial use.\n\n## Publisher:\n\n[adelpro](https://clawhub.ai/user/adelpro)\n\n### License/Terms of Use:\n\nMIT-0\n\n## Use Case:\n\nDevelopers and agent builders use this skill to analyze OpenClaw session history, identify recurring behavior and error patterns, and generate reviewed learning artifacts or optimization suggestions for self-improving agents.\n\n### Deployment Geography for Use:\n\nGlobal\n\n## Known Risks and Mitigations:\n\nRisk: The skill reads recent OpenClaw session logs and may process private interaction history.\n\nMitigation: Run analysis with an explicit --agent scope when possible, and use --all only when broad review is needed.\n\nRisk: Session-derived snippets may retain sensitive details despite redaction.\n\nMitigation: Review stored learning contents periodically and run the prune command if sensitive data may have appeared in tool errors or session output.\n\nRisk: The skill persists derived instincts, patterns, and optimization suggestions locally.\n\nMitigation: Use the documented retention limits and prune command to remove stored learning data before sharing or retiring a workspace.\n\n## Reference(s):\n\n- [ClawHub Skill Page](https://clawhub.ai/adelpro/skills/continue-learning)\n\n## Skill Output:\n\n**Output Type(s):** [text, JSON files, shell commands, guidance]\n\n**Output Format:** [Console text plus local JSON and JSONL learning files]\n\n**Output Parameters:** [1D]\n\n**Other Properties Related to Output:** [Requires Node.js; analysis requires an explicit --agent scope or explicit --all selection.]\n\n## Skill Version(s):\n\n1.3.2 (source: server release metadata and SKILL.md frontmatter)\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 v1.3.1: 5 files, 8725 bytes\n\nFiles: README.md (237b), scripts/analyze.mjs (12927b), skill-card.md (2085b), SKILL.md (7682b), _meta.json (136b)\n\nFile v1.3.1:SKILL.md\n\n---\nname: continue-learning\nslug: continue-learning\nversion: 1.3.1\ndescription: |\n  Instinct-based learning system for OpenClaw. Analyzes sessions, detects patterns, creates atomic learnings with confidence scoring, and suggests optimizations for self-evolution.\n  \n  Works alongside agent-self-improvement for complete learning: internal session analysis + external user feedback.\n  \n  Use when: you want your AI agent to learn from its own behavior, improve over time, discover optimization opportunities, or build a self-improving automation system.\n  \n  Don't use when: static agent behavior is preferred.\ntriggers:\n  - continuous learning\n  - self improving agent\n  - agent evolution\n  - pattern detection\n  - session analysis\n  - ai learning\n  - agent optimization\n  - automation improvement\n  - self evolution\nmetadata:\n  openclaw:\n    emoji: \"🧠\"\n    requires:\n      bins: [\"node\"]\n---\n\n# Continuous Learning for AI Agents\n\nAn instinct-based learning system that helps AI agents improve themselves through observation and pattern detection.\n\n## What This Skill Does\n\n- **Analyzes session history** - Reviews agent interactions and outputs\n- **Detects patterns** - Identifies recurring behaviors, preferences, workflows\n- **Creates instincts** - Atomic learnings with confidence scores\n- **Suggests optimizations** - Based on observed behavior patterns\n- **Enables self-evolution** - Converts insights into improvements\n\n## When to Use\n\n**Use when:**\n- Building self-improving AI agents\n- Want agent to learn from interactions\n- Discovering optimization opportunities\n- Creating adaptive automation\n- Tracking behavioral patterns\n\n**Skip when:**\n- Static, unchanging behavior preferred\n- No session history available\n- Simple, deterministic workflows only\n\n## Architecture\n\n```\n~/.openclaw/agents/ (session .jsonl files)\n        │\n        ▼\n┌───────────────────────────────────────────┐\n│ analyze.mjs                                │\n│ • Reads session history                   │\n│ • Extracts tool calls & errors             │\n│ • Detects patterns                         │\n└───────────────────────────────────────────┘\n        │\n        ▼\n┌───────────────────────────────────────────┐\n│ memory/learning/                           │\n│ • instincts.jsonl (atomic learnings)       │\n│ • patterns.json (aggregated)              │\n│ • optimizations.json (suggestions)         │\n└───────────────────────────────────────────┘\n```\n\n## External Feedback (Sub-Skill)\n\nThis skill works with **agent-self-improvement** (ClawHub) for external user feedback capture:\n\n- **Internal Learning**: Session analysis (this skill)\n- **External Learning**: User feedback via `SKILL:agent-self-improvement`\n\n### Combined Usage\n\n```\n# Nightly: Internal analysis\nSKILL:openclaw-continuous-learning --analyze\n\n# After any output: Capture feedback\nSKILL:agent-self-improvement --job <task> --feedback \"<user response>\"\n\n# Daily: Generate combined improvements\nSKILL:agent-self-improvement --improve all\n```\n\n### Feedback Flow\n\n```\nUser Response → agent-self-improvement → Directive Hints\n        ↓\nSession Analysis → openclaw-continuous-learning → Internal Patterns\n        ↓\nCombined Insights → Agent Optimization\n```\n\nBoth skills store learnings in `memory/learning/` and can reference each other's data.\n\n## Confidence Scoring\n\n| Score | Meaning | Behavior |\n|-------|---------|----------|\n| 0.3 | Tentative | Suggested but not enforced |\n| 0.5 | Moderate | Applied when relevant |\n| 0.7 | Strong | Auto-approved |\n| 0.9 | Core behavior | Always apply |\n\n**Confidence increases when:**\n- Pattern observed repeatedly\n- User doesn't correct behavior\n- Multiple observations agree\n\n**Confidence decreases when:**\n- User explicitly corrects\n- Pattern not observed recently\n- Contradicting evidence appears\n\n## Key Concepts\n\n### Instincts\n\nAn instinct is a small learned behavior:\n\n```yaml\nid: prefer-simplicity\ntrigger: \"when solving problems\"\nconfidence: 0.7\ndomain: problem_solving\n---\n# Prefer Simple Solutions\n\n## Action\nAlways choose the simplest solution that meets requirements.\n\n## Evidence\n- Observed preference for minimal code\n- User corrected over-engineered approaches\n```\n\n### Patterns\n\nAggregated observations grouped by category:\n- code_style\n- testing\n- git\n- debugging\n- workflow\n- communication\n\n### Optimizations\n\nActionable improvements derived from patterns.\n\n## Use Cases\n\n### 1. Agent Self-Improvement\n\n```\nAgent observes its own sessions:\n- What works consistently?\n- What gets corrected?\n- What patterns emerge?\n\nCreates instincts → Applies high-confidence patterns\n```\n\n### 2. User Preference Learning\n\n```\nLearn user preferences from interactions:\n- Coding style preferences\n- Communication preferences\n- Workflow preferences\n\nAdapt behavior accordingly\n```\n\n### 3. Performance Optimization\n\n```\nDetect performance patterns:\n- Slow operations\n- Bottlenecks\n- Optimization opportunities\n\nSuggest improvements\n```\n\n### 4. Error Pattern Detection\n\n```\nTrack error patterns:\n- Common failures\n- Resolution strategies\n- Prevention approaches\n\nBuild error-handling instincts\n```\n\n## Quick Start\n\n```bash\n# Analyze sessions (reads agent .jsonl files from ~/.openclaw/agents/)\ncd ~/.openclaw/workspace/skills/openclaw-continuous-learning\nnode scripts/analyze.mjs\n\n# List learned instincts\nnode scripts/analyze.mjs instincts\n\n# Show optimizations\nnode scripts/analyze.mjs list\n\n# Show error patterns\nnode scripts/analyze.mjs errors\n```\n\n## Setup\n\n### 1. Create storage directory\n\n```bash\nmkdir -p ~/.openclaw/workspace/memory/learning\n```\n\n### 2. Schedule analysis\n\nAdd to cron for periodic analysis:\n\n```json\n{\n  \"id\": \"continuous-learning\",\n  \"schedule\": \"0 22 * * *\"\n}\n```\n\n### 3. Integrate with daily tips\n\nConnect to daily summary for optimization delivery.\n\n## File Structure\n\n```\n~/.openclaw/workspace/\n└── memory/\n    └── learning/\n        ├── instincts.jsonl    # Atomic learnings\n        ├── patterns.json      # Aggregated patterns\n        └── optimizations.json # Suggestions\n```\n\n## Example Output\n\n```\n🧠 Learning Report\n\nPatterns Detected:\n- prefer-simplicity (0.7) ↑2\n- test-first (0.5) ↑1\n- commit-often (0.3) new\n\nConfidence Changes:\n- minimal-code: 0.5 → 0.7\n\nSuggested:\n1. Prioritize simple solutions\n2. Add pre-commit hooks\n3. Enable stricter typing\n```\n\n## Best Practices\n\n1. **Start simple** - Few patterns, low confidence\n2. **Validate often** - Check if patterns still hold\n3. **Review suggestions** - Don't auto-apply everything\n4. **Track confidence** - Update based on results\n5. **Export/share** - Build library of common patterns\n\n## FAQ\n\n**How is this different from memory?**\nMemory stores facts. This learns behavioral patterns and preferences.\n\n**How long to see results?**\nDepends on session volume. Typically 1-2 weeks for meaningful patterns.\n\n**Is it safe to auto-apply?**\nOnly high-confidence (0.7+) patterns. Always review suggestions first.\n\n## Related Skills\n\n- **skill-engineer** - Quality-gated skill development\n- **compound-engineering** - Session review and learning\n- **memory-setup** - Memory configuration\n- **openclaw-daily-tips** - Daily optimization tips\n\n---\n\n**Version:** 1.1.0  \n**Inspired by:** Anthropic's continuous learning patterns, Claude Code homunculus\n\nFile v1.3.1:README.md\n\n# continue-learning\n\nOpenClaw instinct-based learning system: session analysis, pattern detection, atomic learnings with confidence scoring, and optimization suggestions.\n\n## Install\n\n```bash\nnpx skills add adelpro/continue-learning\n```\n\nFile v1.3.1:_meta.json\n\n{\n  \"ownerId\": \"kn72cbvk9f5n48msm4t4sj3wyh80n2eb\",\n  \"slug\": \"continue-learning\",\n  \"version\": \"1.3.1\",\n  \"publishedAt\": 1787253766898\n}\n\nFile v1.3.1:skill-card.md\n\n## Description:\n\nInstinct-based learning system for OpenClaw that analyzes sessions, detects patterns, creates atomic learnings with confidence scoring, and suggests optimizations for self-evolution.\n\nThis skill is ready for commercial/non-commercial use.\n\n## Publisher:\n\n[adelpro](https://clawhub.ai/user/adelpro)\n\n### License/Terms of Use:\n\nMIT-0\n\n## Use Case:\n\nDevelopers and agent operators use this skill to review OpenClaw session history, identify recurring tool and error patterns, and generate learning artifacts that can guide future agent behavior.\n\n### Deployment Geography for Use:\n\nGlobal\n\n## Known Risks and Mitigations:\n\nRisk: The skill reads local OpenClaw session histories, which may include sensitive interaction content.\n\nMitigation: Use it only on sessions acceptable for retention and avoid running it where secrets or sensitive business data may be present.\n\nRisk: The skill writes persistent learning files without strong privacy, scoping, or deletion controls.\n\nMitigation: Review the memory/learning directory periodically and remove stale or sensitive learnings before reuse or sharing.\n\nRisk: Scheduled analysis can repeatedly process new session history.\n\nMitigation: Enable scheduled analysis only after deciding the intended scope and retention window.\n\n## Reference(s):\n\n- [ClawHub skill page](https://clawhub.ai/adelpro/skills/continue-learning)\n\n## Skill Output:\n\n**Output Type(s):** [text, json, shell commands, configuration, guidance]\n\n**Output Format:** [Console text plus JSON and JSONL learning files]\n\n**Output Parameters:** [1D]\n\n**Other Properties Related to Output:** [Creates or updates instincts.jsonl, patterns.json, and optimizations.json under the OpenClaw workspace learning directory.]\n\n## Skill Version(s):\n\n1.3.1 (source: server evidence and SKILL.md frontmatter)\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 v1.3.0: 4 files, 8571 bytes\n\nFiles: scripts/analyze.mjs (12927b), skill-card.md (2433b), SKILL.md (7704b), _meta.json (136b)\n\nFile v1.3.0:SKILL.md\n\n---\nname: openclaw-continuous-learning\nslug: openclaw-continuous-learning\nversion: 1.2.1\ndescription: |\n  Instinct-based learning system for OpenClaw. Analyzes sessions, detects patterns, creates atomic learnings with confidence scoring, and suggests optimizations for self-evolution.\n  \n  Works alongside agent-self-improvement for complete learning: internal session analysis + external user feedback.\n  \n  Use when: you want your AI agent to learn from its own behavior, improve over time, discover optimization opportunities, or build a self-improving automation system.\n  \n  Don't use when: static agent behavior is preferred.\ntriggers:\n  - continuous learning\n  - self improving agent\n  - agent evolution\n  - pattern detection\n  - session analysis\n  - ai learning\n  - agent optimization\n  - automation improvement\n  - self evolution\nmetadata:\n  openclaw:\n    emoji: \"🧠\"\n    requires:\n      bins: [\"node\"]\n---\n\n# Continuous Learning for AI Agents\n\nAn instinct-based learning system that helps AI agents improve themselves through observation and pattern detection.\n\n## What This Skill Does\n\n- **Analyzes session history** - Reviews agent interactions and outputs\n- **Detects patterns** - Identifies recurring behaviors, preferences, workflows\n- **Creates instincts** - Atomic learnings with confidence scores\n- **Suggests optimizations** - Based on observed behavior patterns\n- **Enables self-evolution** - Converts insights into improvements\n\n## When to Use\n\n**Use when:**\n- Building self-improving AI agents\n- Want agent to learn from interactions\n- Discovering optimization opportunities\n- Creating adaptive automation\n- Tracking behavioral patterns\n\n**Skip when:**\n- Static, unchanging behavior preferred\n- No session history available\n- Simple, deterministic workflows only\n\n## Architecture\n\n```\n~/.openclaw/agents/ (session .jsonl files)\n        │\n        ▼\n┌───────────────────────────────────────────┐\n│ analyze.mjs                                │\n│ • Reads session history                   │\n│ • Extracts tool calls & errors             │\n│ • Detects patterns                         │\n└───────────────────────────────────────────┘\n        │\n        ▼\n┌───────────────────────────────────────────┐\n│ memory/learning/                           │\n│ • instincts.jsonl (atomic learnings)       │\n│ • patterns.json (aggregated)              │\n│ • optimizations.json (suggestions)         │\n└───────────────────────────────────────────┘\n```\n\n## External Feedback (Sub-Skill)\n\nThis skill works with **agent-self-improvement** (ClawHub) for external user feedback capture:\n\n- **Internal Learning**: Session analysis (this skill)\n- **External Learning**: User feedback via `SKILL:agent-self-improvement`\n\n### Combined Usage\n\n```\n# Nightly: Internal analysis\nSKILL:openclaw-continuous-learning --analyze\n\n# After any output: Capture feedback\nSKILL:agent-self-improvement --job <task> --feedback \"<user response>\"\n\n# Daily: Generate combined improvements\nSKILL:agent-self-improvement --improve all\n```\n\n### Feedback Flow\n\n```\nUser Response → agent-self-improvement → Directive Hints\n        ↓\nSession Analysis → openclaw-continuous-learning → Internal Patterns\n        ↓\nCombined Insights → Agent Optimization\n```\n\nBoth skills store learnings in `memory/learning/` and can reference each other's data.\n\n## Confidence Scoring\n\n| Score | Meaning | Behavior |\n|-------|---------|----------|\n| 0.3 | Tentative | Suggested but not enforced |\n| 0.5 | Moderate | Applied when relevant |\n| 0.7 | Strong | Auto-approved |\n| 0.9 | Core behavior | Always apply |\n\n**Confidence increases when:**\n- Pattern observed repeatedly\n- User doesn't correct behavior\n- Multiple observations agree\n\n**Confidence decreases when:**\n- User explicitly corrects\n- Pattern not observed recently\n- Contradicting evidence appears\n\n## Key Concepts\n\n### Instincts\n\nAn instinct is a small learned behavior:\n\n```yaml\nid: prefer-simplicity\ntrigger: \"when solving problems\"\nconfidence: 0.7\ndomain: problem_solving\n---\n# Prefer Simple Solutions\n\n## Action\nAlways choose the simplest solution that meets requirements.\n\n## Evidence\n- Observed preference for minimal code\n- User corrected over-engineered approaches\n```\n\n### Patterns\n\nAggregated observations grouped by category:\n- code_style\n- testing\n- git\n- debugging\n- workflow\n- communication\n\n### Optimizations\n\nActionable improvements derived from patterns.\n\n## Use Cases\n\n### 1. Agent Self-Improvement\n\n```\nAgent observes its own sessions:\n- What works consistently?\n- What gets corrected?\n- What patterns emerge?\n\nCreates instincts → Applies high-confidence patterns\n```\n\n### 2. User Preference Learning\n\n```\nLearn user preferences from interactions:\n- Coding style preferences\n- Communication preferences\n- Workflow preferences\n\nAdapt behavior accordingly\n```\n\n### 3. Performance Optimization\n\n```\nDetect performance patterns:\n- Slow operations\n- Bottlenecks\n- Optimization opportunities\n\nSuggest improvements\n```\n\n### 4. Error Pattern Detection\n\n```\nTrack error patterns:\n- Common failures\n- Resolution strategies\n- Prevention approaches\n\nBuild error-handling instincts\n```\n\n## Quick Start\n\n```bash\n# Analyze sessions (reads agent .jsonl files from ~/.openclaw/agents/)\ncd ~/.openclaw/workspace/skills/openclaw-continuous-learning\nnode scripts/analyze.mjs\n\n# List learned instincts\nnode scripts/analyze.mjs instincts\n\n# Show optimizations\nnode scripts/analyze.mjs list\n\n# Show error patterns\nnode scripts/analyze.mjs errors\n```\n\n## Setup\n\n### 1. Create storage directory\n\n```bash\nmkdir -p ~/.openclaw/workspace/memory/learning\n```\n\n### 2. Schedule analysis\n\nAdd to cron for periodic analysis:\n\n```json\n{\n  \"id\": \"continuous-learning\",\n  \"schedule\": \"0 22 * * *\"\n}\n```\n\n### 3. Integrate with daily tips\n\nConnect to daily summary for optimization delivery.\n\n## File Structure\n\n```\n~/.openclaw/workspace/\n└── memory/\n    └── learning/\n        ├── instincts.jsonl    # Atomic learnings\n        ├── patterns.json      # Aggregated patterns\n        └── optimizations.json # Suggestions\n```\n\n## Example Output\n\n```\n🧠 Learning Report\n\nPatterns Detected:\n- prefer-simplicity (0.7) ↑2\n- test-first (0.5) ↑1\n- commit-often (0.3) new\n\nConfidence Changes:\n- minimal-code: 0.5 → 0.7\n\nSuggested:\n1. Prioritize simple solutions\n2. Add pre-commit hooks\n3. Enable stricter typing\n```\n\n## Best Practices\n\n1. **Start simple** - Few patterns, low confidence\n2. **Validate often** - Check if patterns still hold\n3. **Review suggestions** - Don't auto-apply everything\n4. **Track confidence** - Update based on results\n5. **Export/share** - Build library of common patterns\n\n## FAQ\n\n**How is this different from memory?**\nMemory stores facts. This learns behavioral patterns and preferences.\n\n**How long to see results?**\nDepends on session volume. Typically 1-2 weeks for meaningful patterns.\n\n**Is it safe to auto-apply?**\nOnly high-confidence (0.7+) patterns. Always review suggestions first.\n\n## Related Skills\n\n- **skill-engineer** - Quality-gated skill development\n- **compound-engineering** - Session review and learning\n- **memory-setup** - Memory configuration\n- **openclaw-daily-tips** - Daily optimization tips\n\n---\n\n**Version:** 1.1.0  \n**Inspired by:** Anthropic's continuous learning patterns, Claude Code homunculus\n\nFile v1.3.0:_meta.json\n\n{\n  \"ownerId\": \"kn72cbvk9f5n48msm4t4sj3wyh80n2eb\",\n  \"slug\": \"continue-learning\",\n  \"version\": \"1.3.0\",\n  \"publishedAt\": 1773510618226\n}\n\nFile v1.3.0:skill-card.md\n\n## Description: <br>\nInstinct-based learning system for OpenClaw that analyzes sessions, detects patterns, creates atomic learnings with confidence scoring, and suggests optimizations for self-evolution. <br>\n\nThis skill is ready for commercial/non-commercial use. <br>\n\n## Publisher: <br>\n[adelpro](https://clawhub.ai/user/adelpro) <br>\n\n### License/Terms of Use: <br>\nMIT-0 <br>\n\n\n## Use Case: <br>\nDevelopers and agent operators use this skill to analyze prior OpenClaw sessions, identify recurring behaviors and errors, and generate local learning artifacts that can guide future agent optimization. <br>\n\n### Deployment Geography for Use: <br>\nGlobal <br>\n\n## Known Risks and Mitigations: <br>\nRisk: Session logs may contain sensitive user, workflow, or tool-output details that could be summarized into local learning files. <br>\nMitigation: Review or scrub sensitive session logs before running the analyzer, and periodically inspect or delete generated memory/learning files. <br>\nRisk: Optional scheduled analysis can repeatedly process new local sessions and persist derived patterns without a fresh manual review each run. <br>\nMitigation: Avoid cron or other recurring scheduling until the recorded data and retention behavior are understood. <br>\nRisk: Optimization suggestions or learned instincts can be incorrect, stale, or too broad for future tasks. <br>\nMitigation: Review suggested optimizations before applying them and treat confidence scores as advisory rather than automatic approval. <br>\n\n\n## Reference(s): <br>\n- [ClawHub Skill Page](https://clawhub.ai/adelpro/openclaw-continuous-learning) <br>\n\n\n## Skill Output: <br>\n**Output Type(s):** [Analysis, Text, Shell commands, Configuration, Files, Guidance] <br>\n**Output Format:** [Markdown guidance with bash examples, console text, and local JSON/JSONL learning files] <br>\n**Output Parameters:** [1D] <br>\n**Other Properties Related to Output:** [Requires Node.js; reads local OpenClaw session logs from ~/.openclaw/agents and writes local memory/learning artifacts.] <br>\n\n## Skill Version(s): <br>\n1.3.0 (source: server release metadata; artifact frontmatter lists 1.2.1) <br>\n\n## Ethical Considerations: <br>\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. <br>\n\nArchive v1.2.0: 3 files, 5799 bytes\n\nFiles: scripts/analyze.mjs (7014b), SKILL.md (7939b), _meta.json (136b)\n\nFile v1.2.0:SKILL.md\n\n---\nname: openclaw-continuous-learning\nslug: openclaw-continuous-learning\nversion: 1.2.0\ndescription: |\n  Instinct-based learning system for OpenClaw. Analyzes sessions, detects patterns, creates atomic learnings with confidence scoring, and suggests optimizations for self-evolution.\n  \n  Works alongside agent-self-improvement for complete learning: internal session analysis + external user feedback.\n  \n  Use when: you want your AI agent to learn from its own behavior, improve over time, discover optimization opportunities, or build a self-improving automation system.\n  \n  Don't use when: static agent behavior is preferred.\ntriggers:\n  - continuous learning\n  - self improving agent\n  - agent evolution\n  - pattern detection\n  - session analysis\n  - ai learning\n  - agent optimization\n  - automation improvement\n  - self evolution\nmetadata:\n  openclaw:\n    emoji: \"🧠\"\n    requires:\n      bins: [\"node\"]\n---\n\n# Continuous Learning for AI Agents\n\nAn instinct-based learning system that helps AI agents improve themselves through observation and pattern detection.\n\n## What This Skill Does\n\n- **Analyzes session history** - Reviews agent interactions and outputs\n- **Detects patterns** - Identifies recurring behaviors, preferences, workflows\n- **Creates instincts** - Atomic learnings with confidence scores\n- **Suggests optimizations** - Based on observed behavior patterns\n- **Enables self-evolution** - Converts insights into improvements\n\n## When to Use\n\n**Use when:**\n- Building self-improving AI agents\n- Want agent to learn from interactions\n- Discovering optimization opportunities\n- Creating adaptive automation\n- Tracking behavioral patterns\n\n**Skip when:**\n- Static, unchanging behavior preferred\n- No session history available\n- Simple, deterministic workflows only\n\n## Architecture\n\n```\nSession Activity\n │\n ▼\n┌─────────────────────────────────────────┐\n│ Session Analysis                         │\n│ • Read interaction logs                  │\n│ • Detect patterns                       │\n│ • Create instincts                       │\n└─────────────────────────────────────────┘\n │\n ▼\n┌─────────────────────────────────────────┐\n│ Instinct Storage                         │\n│ • instincts.jsonl (atomic learnings)     │\n│ • patterns.json (aggregated)             │\n│ • optimizations.json (suggestions)       │\n└─────────────────────────────────────────┘\n │\n ▼\n┌─────────────────────────────────────────┐\n│ Optimization Delivery                    │\n│ • Daily tips                            │\n│ • Configuration suggestions             │\n│ • Workflow improvements                 │\n└─────────────────────────────────────────┘\n```\n\n## External Feedback (Sub-Skill)\n\nThis skill works with **agent-self-improvement** (ClawHub) for external user feedback capture:\n\n- **Internal Learning**: Session analysis (this skill)\n- **External Learning**: User feedback via `SKILL:agent-self-improvement`\n\n### Combined Usage\n\n```\n# Nightly: Internal analysis\nSKILL:openclaw-continuous-learning --analyze\n\n# After any output: Capture feedback\nSKILL:agent-self-improvement --job <task> --feedback \"<user response>\"\n\n# Daily: Generate combined improvements\nSKILL:agent-self-improvement --improve all\n```\n\n### Feedback Flow\n\n```\nUser Response → agent-self-improvement → Directive Hints\n        ↓\nSession Analysis → openclaw-continuous-learning → Internal Patterns\n        ↓\nCombined Insights → Agent Optimization\n```\n\nBoth skills store learnings in `memory/learning/` and can reference each other's data.\n\n## Confidence Scoring\n\n| Score | Meaning | Behavior |\n|-------|---------|----------|\n| 0.3 | Tentative | Suggested but not enforced |\n| 0.5 | Moderate | Applied when relevant |\n| 0.7 | Strong | Auto-approved |\n| 0.9 | Core behavior | Always apply |\n\n**Confidence increases when:**\n- Pattern observed repeatedly\n- User doesn't correct behavior\n- Multiple observations agree\n\n**Confidence decreases when:**\n- User explicitly corrects\n- Pattern not observed recently\n- Contradicting evidence appears\n\n## Key Concepts\n\n### Instincts\n\nAn instinct is a small learned behavior:\n\n```yaml\nid: prefer-simplicity\ntrigger: \"when solving problems\"\nconfidence: 0.7\ndomain: problem_solving\n---\n# Prefer Simple Solutions\n\n## Action\nAlways choose the simplest solution that meets requirements.\n\n## Evidence\n- Observed preference for minimal code\n- User corrected over-engineered approaches\n```\n\n### Patterns\n\nAggregated observations grouped by category:\n- code_style\n- testing\n- git\n- debugging\n- workflow\n- communication\n\n### Optimizations\n\nActionable improvements derived from patterns.\n\n## Use Cases\n\n### 1. Agent Self-Improvement\n\n```\nAgent observes its own sessions:\n- What works consistently?\n- What gets corrected?\n- What patterns emerge?\n\nCreates instincts → Applies high-confidence patterns\n```\n\n### 2. User Preference Learning\n\n```\nLearn user preferences from interactions:\n- Coding style preferences\n- Communication preferences\n- Workflow preferences\n\nAdapt behavior accordingly\n```\n\n### 3. Performance Optimization\n\n```\nDetect performance patterns:\n- Slow operations\n- Bottlenecks\n- Optimization opportunities\n\nSuggest improvements\n```\n\n### 4. Error Pattern Detection\n\n```\nTrack error patterns:\n- Common failures\n- Resolution strategies\n- Prevention approaches\n\nBuild error-handling instincts\n```\n\n## Quick Start\n\n```bash\n# Analyze sessions\nnode /path/to/scripts/analyze.mjs\n\n# List learned instincts\nnode /path/to/scripts/analyze.mjs instincts\n\n# Show optimizations\nnode /path/to/scripts/analyze.mjs list\n```\n\n## Setup\n\n### 1. Create storage directory\n\n```bash\nmkdir -p ~/.openclaw/workspace/memory/learning\n```\n\n### 2. Schedule analysis\n\nAdd to cron for periodic analysis:\n\n```json\n{\n  \"id\": \"continuous-learning\",\n  \"schedule\": \"0 22 * * *\"\n}\n```\n\n### 3. Integrate with daily tips\n\nConnect to daily summary for optimization delivery.\n\n## File Structure\n\n```\n~/.openclaw/workspace/\n└── memory/\n    └── learning/\n        ├── instincts.jsonl    # Atomic learnings\n        ├── patterns.json      # Aggregated patterns\n        └── optimizations.json # Suggestions\n```\n\n## Example Output\n\n```\n🧠 Learning Report\n\nPatterns Detected:\n- prefer-simplicity (0.7) ↑2\n- test-first (0.5) ↑1\n- commit-often (0.3) new\n\nConfidence Changes:\n- minimal-code: 0.5 → 0.7\n\nSuggested:\n1. Prioritize simple solutions\n2. Add pre-commit hooks\n3. Enable stricter typing\n```\n\n## Best Practices\n\n1. **Start simple** - Few patterns, low confidence\n2. **Validate often** - Check if patterns still hold\n3. **Review suggestions** - Don't auto-apply everything\n4. **Track confidence** - Update based on results\n5. **Export/share** - Build library of common patterns\n\n## FAQ\n\n**How is this different from memory?**\nMemory stores facts. This learns behavioral patterns and preferences.\n\n**How long to see results?**\nDepends on session volume. Typically 1-2 weeks for meaningful patterns.\n\n**Is it safe to auto-apply?**\nOnly high-confidence (0.7+) patterns. Always review suggestions first.\n\n## Related Skills\n\n- **skill-engineer** - Quality-gated skill development\n- **compound-engineering** - Session review and learning\n- **memory-setup** - Memory configuration\n- **openclaw-daily-tips** - Daily optimization tips\n\n---\n\n**Version:** 1.1.0  \n**Inspired by:** Anthropic's continuous learning patterns, Claude Code homunculus\n\nFile v1.2.0:_meta.json\n\n{\n  \"ownerId\": \"kn72cbvk9f5n48msm4t4sj3wyh80n2eb\",\n  \"slug\": \"continue-learning\",\n  \"version\": \"1.2.0\",\n  \"publishedAt\": 1773478829051\n}\n\nArchive v1.1.0: 3 files, 5449 bytes\n\nFiles: scripts/analyze.mjs (7090b), SKILL.md (6950b), _meta.json (136b)\n\nFile v1.1.0:SKILL.md\n\n---\nname: openclaw-continuous-learning\nslug: openclaw-continuous-learning\nversion: 1.1.0\ndescription: |\n  Instinct-based learning system for OpenClaw. Analyzes sessions, detects patterns, creates atomic learnings with confidence scoring, and suggests optimizations for self-evolution.\n  \n  Use when: you want your AI agent to learn from its own behavior, improve over time, discover optimization opportunities, or build a self-improving automation system.\n  \n  Don't use when: static agent behavior is preferred.\ntriggers:\n  - continuous learning\n  - self improving agent\n  - agent evolution\n  - pattern detection\n  - session analysis\n  - ai learning\n  - agent optimization\n  - automation improvement\n  - self evolution\nmetadata:\n  openclaw:\n    emoji: \"🧠\"\n    requires:\n      bins: [\"node\"]\n---\n\n# Continuous Learning for AI Agents\n\nAn instinct-based learning system that helps AI agents improve themselves through observation and pattern detection.\n\n## What This Skill Does\n\n- **Analyzes session history** - Reviews agent interactions and outputs\n- **Detects patterns** - Identifies recurring behaviors, preferences, workflows\n- **Creates instincts** - Atomic learnings with confidence scores\n- **Suggests optimizations** - Based on observed behavior patterns\n- **Enables self-evolution** - Converts insights into improvements\n\n## When to Use\n\n**Use when:**\n- Building self-improving AI agents\n- Want agent to learn from interactions\n- Discovering optimization opportunities\n- Creating adaptive automation\n- Tracking behavioral patterns\n\n**Skip when:**\n- Static, unchanging behavior preferred\n- No session history available\n- Simple, deterministic workflows only\n\n## Architecture\n\n```\nSession Activity\n │\n ▼\n┌─────────────────────────────────────────┐\n│ Session Analysis                         │\n│ • Read interaction logs                  │\n│ • Detect patterns                       │\n│ • Create instincts                       │\n└─────────────────────────────────────────┘\n │\n ▼\n┌─────────────────────────────────────────┐\n│ Instinct Storage                         │\n│ • instincts.jsonl (atomic learnings)     │\n│ • patterns.json (aggregated)             │\n│ • optimizations.json (suggestions)       │\n└─────────────────────────────────────────┘\n │\n ▼\n┌─────────────────────────────────────────┐\n│ Optimization Delivery                    │\n│ • Daily tips                            │\n│ • Configuration suggestions             │\n│ • Workflow improvements                 │\n└─────────────────────────────────────────┘\n```\n\n## Confidence Scoring\n\n| Score | Meaning | Behavior |\n|-------|---------|----------|\n| 0.3 | Tentative | Suggested but not enforced |\n| 0.5 | Moderate | Applied when relevant |\n| 0.7 | Strong | Auto-approved |\n| 0.9 | Core behavior | Always apply |\n\n**Confidence increases when:**\n- Pattern observed repeatedly\n- User doesn't correct behavior\n- Multiple observations agree\n\n**Confidence decreases when:**\n- User explicitly corrects\n- Pattern not observed recently\n- Contradicting evidence appears\n\n## Key Concepts\n\n### Instincts\n\nAn instinct is a small learned behavior:\n\n```yaml\nid: prefer-simplicity\ntrigger: \"when solving problems\"\nconfidence: 0.7\ndomain: problem_solving\n---\n# Prefer Simple Solutions\n\n## Action\nAlways choose the simplest solution that meets requirements.\n\n## Evidence\n- Observed preference for minimal code\n- User corrected over-engineered approaches\n```\n\n### Patterns\n\nAggregated observations grouped by category:\n- code_style\n- testing\n- git\n- debugging\n- workflow\n- communication\n\n### Optimizations\n\nActionable improvements derived from patterns.\n\n## Use Cases\n\n### 1. Agent Self-Improvement\n\n```\nAgent observes its own sessions:\n- What works consistently?\n- What gets corrected?\n- What patterns emerge?\n\nCreates instincts → Applies high-confidence patterns\n```\n\n### 2. User Preference Learning\n\n```\nLearn user preferences from interactions:\n- Coding style preferences\n- Communication preferences\n- Workflow preferences\n\nAdapt behavior accordingly\n```\n\n### 3. Performance Optimization\n\n```\nDetect performance patterns:\n- Slow operations\n- Bottlenecks\n- Optimization opportunities\n\nSuggest improvements\n```\n\n### 4. Error Pattern Detection\n\n```\nTrack error patterns:\n- Common failures\n- Resolution strategies\n- Prevention approaches\n\nBuild error-handling instincts\n```\n\n## Quick Start\n\n```bash\n# Analyze sessions\nnode /path/to/scripts/analyze.mjs\n\n# List learned instincts\nnode /path/to/scripts/analyze.mjs instincts\n\n# Show optimizations\nnode /path/to/scripts/analyze.mjs list\n```\n\n## Setup\n\n### 1. Create storage directory\n\n```bash\nmkdir -p ~/.openclaw/workspace/memory/learning\n```\n\n### 2. Schedule analysis\n\nAdd to cron for periodic analysis:\n\n```json\n{\n  \"id\": \"continuous-learning\",\n  \"schedule\": \"0 22 * * *\"\n}\n```\n\n### 3. Integrate with daily tips\n\nConnect to daily summary for optimization delivery.\n\n## File Structure\n\n```\n~/.openclaw/workspace/\n└── memory/\n    └── learning/\n        ├── instincts.jsonl    # Atomic learnings\n        ├── patterns.json      # Aggregated patterns\n        └── optimizations.json # Suggestions\n```\n\n## Example Output\n\n```\n🧠 Learning Report\n\nPatterns Detected:\n- prefer-simplicity (0.7) ↑2\n- test-first (0.5) ↑1\n- commit-often (0.3) new\n\nConfidence Changes:\n- minimal-code: 0.5 → 0.7\n\nSuggested:\n1. Prioritize simple solutions\n2. Add pre-commit hooks\n3. Enable stricter typing\n```\n\n## Best Practices\n\n1. **Start simple** - Few patterns, low confidence\n2. **Validate often** - Check if patterns still hold\n3. **Review suggestions** - Don't auto-apply everything\n4. **Track confidence** - Update based on results\n5. **Export/share** - Build library of common patterns\n\n## FAQ\n\n**How is this different from memory?**\nMemory stores facts. This learns behavioral patterns and preferences.\n\n**How long to see results?**\nDepends on session volume. Typically 1-2 weeks for meaningful patterns.\n\n**Is it safe to auto-apply?**\nOnly high-confidence (0.7+) patterns. Always review suggestions first.\n\n## Related Skills\n\n- **skill-engineer** - Quality-gated skill development\n- **compound-engineering** - Session review and learning\n- **memory-setup** - Memory configuration\n- **openclaw-daily-tips** - Daily optimization tips\n\n---\n\n**Version:** 1.1.0  \n**Inspired by:** Anthropic's continuous learning patterns, Claude Code homunculus\n\nFile v1.1.0:_meta.json\n\n{\n  \"ownerId\": \"kn72cbvk9f5n48msm4t4sj3wyh80n2eb\",\n  \"slug\": \"continue-learning\",\n  \"version\": \"1.1.0\",\n  \"publishedAt\": 1772011023763\n}","readmeExcerpt":"Skill: continue-learning Owner: adelpro Summary: Instinct-based learning system for OpenClaw. Analyzes sessions, detects patterns, creates atomic learnings with confidence scoring, and suggests optimizations for self-evolution. Works alongside agent-self-improvement for complete learning: internal session analysis + external user feedback. Use when: you want your AI agent to learn from its own behavior, improve over ","codeSnippets":[],"executableExamples":[{"language":"text","snippet":"~/.openclaw/agents/ (session .jsonl files)\n        │\n        ▼\n┌───────────────────────────────────────────┐\n│ analyze.mjs                                │\n│ • Reads session history                   │\n│ • Extracts tool calls & errors             │\n│ • Detects patterns                         │\n└───────────────────────────────────────────┘\n        │\n        ▼\n┌───────────────────────────────────────────┐\n│ memory/learning/                           │\n│ • instincts.jsonl (atomic learnings)       │\n│ • patterns.json (aggregated)              │\n│ • optimizations.json (suggestions)         │\n└───────────────────────────────────────────┘"},{"language":"text","snippet":"# Nightly: Internal analysis\nSKILL:openclaw-continuous-learning --analyze\n\n# After any output: Capture feedback\nSKILL:agent-self-improvement --job <task> --feedback \"<user response>\"\n\n# Daily: Generate combined improvements\nSKILL:agent-self-improvement --improve all"},{"language":"text","snippet":"User Response → agent-self-improvement → Directive Hints\n        ↓\nSession Analysis → openclaw-continuous-learning → Internal Patterns\n        ↓\nCombined Insights → Agent Optimization"},{"language":"yaml","snippet":"id: prefer-simplicity\ntrigger: \"when solving problems\"\nconfidence: 0.7\ndomain: problem_solving\n---\n# Prefer Simple Solutions\n\n## Action\nAlways choose the simplest solution that meets requirements.\n\n## Evidence\n- Observed preference for minimal code\n- User corrected over-engineered approaches"},{"language":"text","snippet":"Agent observes its own sessions:\n- What works consistently?\n- What gets corrected?\n- What patterns emerge?\n\nCreates instincts → Applies high-confidence patterns"},{"language":"text","snippet":"Learn user preferences from interactions:\n- Coding style preferences\n- Communication preferences\n- Workflow preferences\n\nAdapt behavior accordingly"}],"parameters":null,"dependencies":[],"permissions":[],"extractedFiles":[{"path":"SKILL.md","content":"---\nname: continue-learning\nslug: continue-learning\nversion: 1.3.2\ndescription: |\n  Instinct-based learning system for OpenClaw. Analyzes sessions, detects patterns, creates atomic learnings with confidence scoring, and suggests optimizations for self-evolution.\n  \n  Works alongside agent-self-improvement for complete learning: internal session analysis + external user feedback.\n  \n  Use when: you want your AI agent to learn from its own behavior, improve over time, discover optimization opportunities, or build a self-improving automation system.\n  \n  Don't use when: static agent behavior is preferred.\ntriggers:\n  - continuous learning\n  - self improving agent\n  - agent evolution\n  - pattern detection\n  - session analysis\n  - ai learning\n  - agent optimization\n  - automation improvement\n  - self evolution\nmetadata:\n  openclaw:\n    emoji: \"🧠\"\n    requires:\n      bins: [\"node\"]\n---\n\n# Continuous Learning for AI Agents\n\nAn instinct-based learning system that helps AI agents improve themselves through observation and pattern detection.\n\n## What This Skill Does\n\n- **Analyzes session history** - Reviews agent interactions and outputs\n- **Detects patterns** - Identifies recurring behaviors, preferences, workflows\n- **Creates instincts** - Atomic learnings with confidence scores\n- **Suggests optimizations** - Based on observed behavior patterns\n- **Enables self-evolution** - Converts insights into improvements\n\n## When to Use\n\n**Use when:**\n- Building self-improving AI agents\n- Want agent to learn from interactions\n- Discovering optimization opportunities\n- Creating adaptive automation\n- Tracking behavioral patterns\n\n**Skip when:**\n- Static, unchanging behavior preferred\n- No session history available\n- Simple, deterministic workflows only\n\n## Architecture\n\n```\n~/.openclaw/agents/ (session .jsonl files)\n        │\n        ▼\n┌───────────────────────────────────────────┐\n│ analyze.mjs                                │\n│ • Reads session history                   │\n│ • Extracts tool calls & errors             │\n│ • Detects patterns                         │\n└───────────────────────────────────────────┘\n        │\n        ▼\n┌───────────────────────────────────────────┐\n│ memory/learning/                           │\n│ • instincts.jsonl (atomic learnings)       │\n│ • patterns.json (aggregated)              │\n│ • optimizations.json (suggestions)         │\n└───────────────────────────────────────────┘\n```\n\n## External Feedback (Sub-Skill)\n\nThis skill works with **agent-self-improvement** (ClawHub) for external user feedback capture:\n\n- **Internal Learning**: Session analysis (this skill)\n- **External Learning**: User feedback via `SKILL:agent-self-improvement`\n\n### Combined Usage\n\n```\n# Nightly: Internal analysis\nSKILL:openclaw-continuous-learning --analyze\n\n# After any output: Capture feedback\nSKILL:agent-self-improvement --job <task> --feedback \"<user response>\"\n\n# Daily: Generate combined improvements\nSKILL:agent-self-improvement --improve all\n```\n\n### Feedback "},{"path":"README.md","content":"# continue-learning\n\nOpenClaw instinct-based learning system: session analysis, pattern detection, atomic learnings with confidence scoring, and optimization suggestions.\n\n## Install\n\n```bash\nnpx skills add adelpro/continue-learning\n```"},{"path":"_meta.json","content":"{\n  \"ownerId\": \"kn72cbvk9f5n48msm4t4sj3wyh80n2eb\",\n  \"slug\": \"continue-learning\",\n  \"version\": \"1.3.2\",\n  \"publishedAt\": 1787254873454\n}"},{"path":"skill-card.md","content":"## Description:\n\nInstinct-based learning system for OpenClaw that analyzes sessions, detects patterns, creates atomic learnings with confidence scoring, and suggests optimizations for self-evolution.\n\nThis skill is ready for commercial/non-commercial use.\n\n## Publisher:\n\n[adelpro](https://clawhub.ai/user/adelpro)\n\n### License/Terms of Use:\n\nMIT-0\n\n## Use Case:\n\nDevelopers and agent builders use this skill to analyze OpenClaw session history, identify recurring behavior and error patterns, and generate reviewed learning artifacts or optimization suggestions for self-improving agents.\n\n### Deployment Geography for Use:\n\nGlobal\n\n## Known Risks and Mitigations:\n\nRisk: The skill reads recent OpenClaw session logs and may process private interaction history.\n\nMitigation: Run analysis with an explicit --agent scope when possible, and use --all only when broad review is needed.\n\nRisk: Session-derived snippets may retain sensitive details despite redaction.\n\nMitigation: Review stored learning contents periodically and run the prune command if sensitive data may have appeared in tool errors or session output.\n\nRisk: The skill persists derived instincts, patterns, and optimization suggestions locally.\n\nMitigation: Use the documented retention limits and prune command to remove stored learning data before sharing or retiring a workspace.\n\n## Reference(s):\n\n- [ClawHub Skill Page](https://clawhub.ai/adelpro/skills/continue-learning)\n\n## Skill Output:\n\n**Output Type(s):** [text, JSON files, shell commands, guidance]\n\n**Output Format:** [Console text plus local JSON and JSONL learning files]\n\n**Output Parameters:** [1D]\n\n**Other Properties Related to Output:** [Requires Node.js; analysis requires an explicit --agent scope or explicit --all selection.]\n\n## Skill Version(s):\n\n1.3.2 (source: server release metadata and SKILL.md frontmatter)\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":1229,"uniquenessScore":44,"reasons":["uniqueness-below-45"]}},"media":{"evidence":{"source":"no-media","verified":false,"confidence":"low","updatedAt":"2026-10-10T09:10:22.804Z","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":"2026-10-10T09:10:22.804Z","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":"2026-10-10T11:49:38.740Z","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"}]}}}