{"id":"59466e8e-a0e8-444c-8db5-6532f8e32ba2","entityType":"agent","slug":"clawhub-testman2025-weekly-review","name":"Weekly Review","canonicalUrl":"https://www.xpersona.co/agent/clawhub-testman2025-weekly-review","canonicalPath":"/agent/clawhub-testman2025-weekly-review","generatedAt":"2026-10-11T17:43:31.300Z","source":"CLAWHUB","claimStatus":"UNCLAIMED","verificationTier":"NONE","summary":{"evidence":{"source":"editorial-content","verified":true,"confidence":"high","updatedAt":"2026-10-11T14:45:34.190Z","emptyReason":null},"description":"面向任意 AI Agent 自动复盘的 Skill（weekly-review / 复盘 / 提示词优化 / 会话清理）。 核心能力覆盖：AI 用量与时间看板、提示词复盘与改写建议、高效/低效归因、 时间分布与 Token/归因趋势图、开放会话对齐清理、定时自动周报。 当用户任务涉及周度复盘、用量查看、提示词改进... Skill: Weekly Review Owner: testman2025 Summary: 面向任意 AI Agent 自动复盘的 Skill（weekly-review / 复盘 / 提示词优化 / 会话清理）。 核心能力覆盖：AI 用量与时间看板、提示词复盘与改写建议、高效/低效归因、 时间分布与 Token/归因趋势图、开放会话对齐清理、定时自动周报。 当用户任务涉及周度复盘、用量查看、提示词改进... Tags: agent-agnostic:1.1.2, agent-fed:1.2.3, automation:1.1.0, latest:1.2.5, office:1.1.0, productivity:1.2.1, retrospective:1.2.3, weekly-review:1.2.3 Version history: v1.2.5 | 2026-07-24T03:31:39.454Z | user","descriptionLabel":"Technical summary","evidenceSummary":"Capability contract not published. No trust telemetry is available yet. 1K downloads reported by the source. Last updated 10/11/2026.","installCommand":"clawhub skill install s17byqgfxb0xcxd564d2gh3yys84v9aa:weekly-review","sourceUrl":"https://clawhub.ai/testman2025/weekly-review","homepage":"https://clawhub.ai/testman2025/skills/weekly-review","primaryLinks":[{"label":"View on ClawHub","url":"https://clawhub.ai/testman2025/weekly-review","kind":"source"},{"label":"Homepage","url":"https://clawhub.ai/testman2025/skills/weekly-review","kind":"homepage"}],"safetyScore":84,"overallRank":62,"popularityScore":60,"trustScore":null,"claimedByName":null,"isOwner":false,"seoDescription":"面向任意 AI Agent 自动复盘的 Skill（weekly-review / 复盘 / 提示词优化 / 会话清理）。 核心能力覆盖：AI 用量与时间看板、提示词复盘与改写建议、高效/低效归因、 时间分布与 Token/归因趋势图、开放会话对齐清理、定时自动周报。 当用户任务涉及周度复盘、用量查看、提示词改进..."},"coverage":{"evidence":{"source":"public-profile","verified":false,"confidence":"medium","updatedAt":"2026-10-11T14:45:34.190Z","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":"2026-10-11T14:45:34.190Z","emptyReason":null},"stars":null,"forks":null,"downloads":1047,"packageName":null,"latestVersion":"1.2.5","tractionLabel":"1K downloads"},"release":{"evidence":{"source":"CLAWHUB","verified":false,"confidence":"medium","updatedAt":"2026-10-11T14:45:34.177Z","emptyReason":null},"lastUpdatedAt":"2026-10-11T14:45:34.190Z","lastCrawledAt":"2026-10-11T14:45:34.177Z","lastIndexedAt":null,"nextCrawlAt":"2026-10-12T14:45:34.177Z","lastVerifiedAt":null,"highlights":[{"version":"1.2.5","createdAt":"2026-07-24T03:31:39.454Z","changelog":"Discoverability: concise Chinese description for prompts, usage, charts, session cleanup, auto review; shorter read_when.","fileCount":13,"zipByteSize":24780},{"version":"1.2.4","createdAt":"2026-07-24T03:29:33.368Z","changelog":"Discoverability: concise Chinese description for prompts, usage, charts, session cleanup, auto review; shorter read_when.","fileCount":13,"zipByteSize":24802},{"version":"1.2.3","createdAt":"2026-07-21T05:28:07.772Z","changelog":"Align report to six-chapter template; horizontal time chart + dual attribution/Token SVG; sanitized public demo.","fileCount":13,"zipByteSize":24669},{"version":"1.2.1","createdAt":"2026-07-21T03:08:51.364Z","changelog":"Bump 1.2.1: sync with GitHub main (Agent-fed review-input path).","fileCount":12,"zipByteSize":17060},{"version":"1.2.0","createdAt":"2026-07-21T02:34:19.113Z","changelog":"Agent-fed weekly review: host Agent collects session facts; skill defines structure and renders Markdown from review-input JSON. SQLite reader moved to legacy/.","fileCount":12,"zipByteSize":16813},{"version":"1.1.2","createdAt":"2026-07-21T02:01:20.707Z","changelog":"Clarify agent-agnostic: any Agent runtime + any compatible local session SQLite (not WorkBuddy-only). Optional usage/automations tables; WEEKLY_REVIEW_DB env.","fileCount":9,"zipByteSize":19403},{"version":"1.1.1","createdAt":"2026-07-20T10:30:44.875Z","changelog":"ClawHub-ready: openclaw metadata (requires python), ClawHub install docs, safety notes.","fileCount":9,"zipByteSize":17706},{"version":"1.1.0","createdAt":"2026-07-20T10:21:34.720Z","changelog":"Flatten to single layer: code at skill root (python -m cli / mcp_server), add README, drop pip-install, stdlib-only","fileCount":9,"zipByteSize":16734}]},"execution":{"evidence":{"source":"CLAWHUB","verified":false,"confidence":"low","updatedAt":null,"emptyReason":"No published capability contract is available yet."},"installCommand":"clawhub skill install s17byqgfxb0xcxd564d2gh3yys84v9aa:weekly-review","setupComplexity":"low","setupSteps":["Setup complexity is classified as HIGH. You must provision dedicated cloud infrastructure or an isolated VM. Do not run this directly on your local workstation.","Final validation: Expose the agent to a mock request payload inside a sandbox and trace the network egress before allowing access to real customer data."],"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-testman2025-weekly-review/snapshot","contractUrl":"https://www.xpersona.co/api/v1/agents/clawhub-testman2025-weekly-review/contract","trustUrl":"https://www.xpersona.co/api/v1/agents/clawhub-testman2025-weekly-review/trust"},"curlExamples":["curl -s \"https://www.xpersona.co/api/v1/agents/clawhub-testman2025-weekly-review/snapshot\"","curl -s \"https://www.xpersona.co/api/v1/agents/clawhub-testman2025-weekly-review/contract\"","curl -s \"https://www.xpersona.co/api/v1/agents/clawhub-testman2025-weekly-review/trust\""],"jsonRequestTemplate":{"query":"summarize this repo","constraints":{"maxLatencyMs":2000,"protocolPreference":["OPENCLEW"]}},"jsonResponseTemplate":{"ok":true,"result":{"summary":"...","confidence":0.9},"meta":{"source":"CLAWHUB","generatedAt":"2026-10-11T17:43:31.297Z"}},"retryPolicy":{"maxAttempts":3,"backoffMs":[500,1500,3500],"retryableConditions":["HTTP_429","HTTP_503","NETWORK_TIMEOUT"]}},"endpoints":{"dossierUrl":"https://www.xpersona.co/api/v1/agents/clawhub-testman2025-weekly-review/dossier","snapshotUrl":"https://www.xpersona.co/api/v1/agents/clawhub-testman2025-weekly-review/snapshot","contractUrl":"https://www.xpersona.co/api/v1/agents/clawhub-testman2025-weekly-review/contract","trustUrl":"https://www.xpersona.co/api/v1/agents/clawhub-testman2025-weekly-review/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":"high","updatedAt":"2026-10-11T14:45:34.190Z","emptyReason":null},"readme":"Skill: Weekly Review\n\nOwner: testman2025\n\nSummary: 面向任意 AI Agent 自动复盘的 Skill（weekly-review / 复盘 / 提示词优化 / 会话清理）。 核心能力覆盖：AI 用量与时间看板、提示词复盘与改写建议、高效/低效归因、 时间分布与 Token/归因趋势图、开放会话对齐清理、定时自动周报。 当用户任务涉及周度复盘、用量查看、提示词改进...\n\nTags: agent-agnostic:1.1.2, agent-fed:1.2.3, automation:1.1.0, latest:1.2.5, office:1.1.0, productivity:1.2.1, retrospective:1.2.3, weekly-review:1.2.3\n\nVersion history:\n\nv1.2.5 | 2026-07-24T03:31:39.454Z | user\n\nDiscoverability: concise Chinese description for prompts, usage, charts, session cleanup, auto review; shorter read_when.\n\nv1.2.4 | 2026-07-24T03:29:33.368Z | user\n\nDiscoverability: concise Chinese description for prompts, usage, charts, session cleanup, auto review; shorter read_when.\n\nv1.2.3 | 2026-07-21T05:28:07.772Z | user\n\nAlign report to six-chapter template; horizontal time chart + dual attribution/Token SVG; sanitized public demo.\n\nv1.2.1 | 2026-07-21T03:08:51.364Z | user\n\nBump 1.2.1: sync with GitHub main (Agent-fed review-input path).\n\nv1.2.0 | 2026-07-21T02:34:19.113Z | user\n\nAgent-fed weekly review: host Agent collects session facts; skill defines structure and renders Markdown from review-input JSON. SQLite reader moved to legacy/.\n\nv1.1.2 | 2026-07-21T02:01:20.707Z | user\n\nClarify agent-agnostic: any Agent runtime + any compatible local session SQLite (not WorkBuddy-only). Optional usage/automations tables; WEEKLY_REVIEW_DB env.\n\nv1.1.1 | 2026-07-20T10:30:44.875Z | user\n\nClawHub-ready: openclaw metadata (requires python), ClawHub install docs, safety notes.\n\nv1.1.0 | 2026-07-20T10:21:34.720Z | user\n\nFlatten to single layer: code at skill root (python -m cli / mcp_server), add README, drop pip-install, stdlib-only\n\nv1.0.1 | 2026-07-20T10:06:17.250Z | user\n\nSelf-contained: bundle weekly_review package (stdlib-only), drop global pip-install, use python -m\n\nv1.0.0 | 2026-07-20T09:58:15.363Z | user\n\nInitial release: data-driven weekly retrospective skill with root-cause triage and action ledger\n\nArchive index:\n\nArchive v1.2.5: 13 files, 24780 bytes\n\nFiles: charts.py (13406b), cli.py (3373b), legacy/analyzer.py (10570b), legacy/README.md (312b), mcp_server.py (3991b), README.md (998b), report.py (15588b), schema/README.md (1652b), schema/review-input.example.json (6932b), skill-card.md (2325b), SKILL.md (2913b), utils.py (506b), _meta.json (132b)\n\nFile v1.2.5:SKILL.md\n\n---\nname: weekly-review\ndescription: >\n 面向任意 AI Agent 自动复盘的 Skill（weekly-review / 复盘 / 提示词优化 / 会话清理）。\n\n 核心能力覆盖：AI 用量与时间看板、提示词复盘与改写建议、高效/低效归因、\n 时间分布与 Token/归因趋势图、开放会话对齐清理、定时自动周报。\n\n 当用户任务涉及周度复盘、用量查看、提示词改进时使用，包括但不限于：\n Token 花了多少、提示词优化改进、一周总结、工作复盘、复盘可视化报告、\n 清理对话、AI 自动复盘。\nversion: 1.2.5\ncategory: 办公效率\nread_when:\n - 提示词改进\n - 本周 AI 用量看板\n - 高低效归因\n - 复盘可视化报告\n - 清理开放对话\n - 定时/自动周报\n - weekly-review / 复盘 / /weekly-review\nmetadata:\n openclaw:\n requires:\n bins:\n - python\n emoji: \"📊\"\n homepage: https://github.com/testman2025/weekly-review-skill\n os:\n - macos\n - linux\n - windows\n---\n\n# AI 用量与提示词复盘助手（weekly-review）\n\n用来做：提示词改进、本周 AI 用量看板、高低效归因、复盘可视化图表、开放会话清理、定时/自动周报。宿主 Agent 采数，本 skill 按六章模板出报告。\n\n## 输出结构（锁定）\n\n1. **一页看板**：`| 指标 | 数值 | 口径说明 |` + **一句话结论** + 可选「辅助图表」PNG 路径 \n2. **分项目分析**：`### 2.x 主题（时长/会话）` + `| 维度 | 内容 |`（做了什么 / DB cwd / 磁盘核对 / 低效 / 提示词）；其他用简表 \n3. **问题清单 + 根因三类归因** + **做得好的（正面范本）** \n4. **本周动作台账 ★**：当场已改 / 观察 / 待落实（或待落实→执行结果） \n5. **待对齐开放会话** \n6. **自动化概览** \n7. **事实更正**（可选，有核对推翻时必写）\n\n**图表约定**：渲染器生成两张定稿样式图——横向「时间投入分布」、双栏「低效归因 + Token 趋势」（SVG）。数据来自 `charts_data`。看板主表仍是 `| 指标 | 数值 | 口径说明 |`。\n\n## 输入\n\n见 `schema/review-input.example.json` 与 `schema/README.md`。\n\n## 操作流程\n\n1. 安装本 skill（见仓库根 README：`npx skills add` / ClawHub / clone）。 \n2. Agent 采集会话/用量，**Glob/ls 核对磁盘**后再归因。 \n3. 填 `review-input`（或对话等价结构）；需要图时先出 PNG 再填 `dashboard.charts`。 \n4. 按六章输出，或 `python -m cli --input review-input.json -o YYYY-MM-DD周度复盘.md`。 \n5. 对开放会话当场拍板。定时场景见仓库 `automations/`。\n\n### CLI\n\n```bash\ncd {SKILL_DIR}\npython -m cli --input schema/review-input.example.json -o 周度复盘.md\n# 图表默认写到报告同目录；关闭：加 --no-charts\n```\n\n### MCP\n\n工具 `run_weekly_review`，参数 `review_input`（对象）。\n\nFile v1.2.5:legacy/README.md\n\n# Legacy（非公开主路径）\n\n此处保留旧版「直接读 SQLite 会话库」分析器，仅供对照或迁移。\n\n**公开能力请使用：**\n\n1. Agent 按 `SKILL.md` 采集本平台数据 \n2. 填入 `schema/review-input.example.json` 结构 \n3. `python -m cli --input review-input.json -o report.md`\n\nFile v1.2.5:README.md\n\n# AI 用量与提示词复盘助手（weekly-review）\n\n面向任意 AI Agent 的自动复盘 Skill。\n\n## 解决什么问题\n\n1. 提升提示词表达能力 \n2. 查看本周 AI 用量 \n3. 总结工作低效与高效 \n4. 复盘可视化图表（时间分布 / 归因 / Token） \n5. 清理/对齐开放对话 \n6. AI 自动/定时周复盘 \n\n## Install\n\n```bash\nnpx skills add testman2025/weekly-review-skill --skill weekly-review\n# or\nopenclaw skills install weekly-review\n```\n\nSee the repo root [README](../../README.md).\n\n## Output shape (locked)\n\n1. Dashboard table (`指标 | 数值 | 口径说明`) + one-liner + optional chart paths \n2. Theme project sections with dimension tables (incl. disk verification) \n3. Problems + three-class root cause + positives \n4. Action ledger \n5. Open sessions \n6. Automations overview \n7. Fact corrections (optional)\n\n## Render\n\n```bash\npython -m cli --input schema/review-input.example.json -o weekly-report.md\n```\n\nSee `SKILL.md` and `schema/`.\n\nFile v1.2.5:schema/README.md\n\n# review-input 约定\n\n对齐定稿模板：`# YYYY-MM-DD 周度复盘（weekly-review skill · 六章结构）`。\n\nAgent 采集事实后填本结构；skill 只渲染结构。**图表**用 `dashboard.charts` 引用已生成的 PNG 路径（辅助图表），不以 Mermaid 为主。\n\n## 图表\n\n渲染默认生成两张**定稿样式 SVG**（标准库）：\n\n1. `chart-时间分布-*.svg` — 横向条形「本周时间投入分布」\n2. `chart-归因Token-*.svg` — 左右双柱「低效条目归因」+「Token 消耗趋势」\n\n数据优先读 `charts_data`：\n\n```json\n\"charts_data\": {\n \"time_distribution\": [{ \"label\": \"社媒 Agent 主题\", \"hours\": 15.4 }],\n \"attribution\": [{ \"label\": \"【用户】指令模糊\", \"count\": 3 }],\n \"token_trend\": [{ \"label\": \"本周\", \"millions\": 2.74 }]\n}\n```\n\n报告用 `![...](path)` 嵌入。`--no-charts` 可关闭。\n\n\n| 字段 | 说明 |\n|------|------|\n| `period` | `start` / `end`（必填） |\n| `meta` | `title` / `methodology` / `version_note` / `task_split` |\n| `source.note` | 数据来源一行说明 |\n| `dashboard.metrics[]` | `{name,value,note}` 看板表 |\n| `dashboard.summary` | 一句话结论 |\n| `dashboard.charts[]` | 辅助图 PNG 相对路径 |\n| `projects[]` | `{title, dimensions{做了什么,…}}` 主题分节 |\n| `other_projects[]` | `{name,hours,note}` |\n| `problems[]` / `positives[]` | 问题表 + 正面范本 |\n| `actions` | `done` / `observing` / `pending` |\n| `open_sessions[]` | 待对齐会话 |\n| `automations.narrative` | 自动化概览（条目列表或长文） |\n| `corrections*` | 可选第七章事实更正 |\n\n完整示例：`review-input.example.json`。\n\nFile v1.2.5:_meta.json\n\n{\n \"ownerId\": \"kn71svn18t9cb73emms0vf78s984tb7p\",\n \"slug\": \"weekly-review\",\n \"version\": \"1.2.5\",\n \"publishedAt\": 1784863899454\n}\n\nFile v1.2.5:skill-card.md\n\n## Description:\n\nWeekly Review helps an agent turn selected weekly AI usage, prompt quality, project, action, open-session, and automation facts into a structured review report with optional charts.\n\nThis skill is ready for commercial/non-commercial use.\n\n## Publisher:\n\n[testman2025](https://clawhub.ai/user/testman2025)\n\n### License/Terms of Use:\n\nMIT-0\n\n## Use Case:\n\nExternal users and developers use this skill to produce weekly AI-usage retrospectives, prompt-improvement notes, inefficiency analysis, action ledgers, open-session cleanup lists, and automation summaries from agent-collected review-input data.\n\n### Deployment Geography for Use:\n\nGlobal\n\n## Known Risks and Mitigations:\n\nRisk: The skill can lead an agent to collect AI session data and inspect local workspace folders without clear built-in scope limits.\n\nMitigation: Use it only for intentionally selected sessions and workspaces, and prefer a user-supplied review-input JSON file.\n\nRisk: The legacy analyzer can read a SQLite session database when explicitly pointed at one.\n\nMitigation: Use the documented public path where the agent supplies structured review-input data, and avoid scanning arbitrary session history.\n\nRisk: Server-resolved import provenance is unavailable for this release.\n\nMitigation: Pin or verify the installer source before running npx or other install commands.\n\n## Reference(s):\n\n- [ClawHub skill page](https://clawhub.ai/testman2025/skills/weekly-review)\n- [Project homepage](https://github.com/testman2025/weekly-review-skill)\n- [Review input schema](schema/README.md)\n\n## Skill Output:\n\n**Output Type(s):** [text, markdown, files, shell commands, configuration, guidance]\n\n**Output Format:** [Markdown report with tables and optional SVG chart files; JSON can be returned when requested.]\n\n**Output Parameters:** [1D]\n\n**Other Properties Related to Output:** [Requires user- or agent-supplied review-input JSON; chart files are generated locally from charts_data.]\n\n## Skill Version(s):\n\n1.2.5 (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\nFile v1.2.5:schema/review-input.example.json\n\n{\n \"period\": {\n \"start\": \"2026-07-13\",\n \"end\": \"2026-07-19\"\n },\n \"meta\": {\n \"title\": \"2026-07-19 周度复盘（weekly-review skill · 六章结构）\",\n \"methodology\": \"本版按 `weekly-review` skill 六章固定结构组织；归因遵循三类框架（思路/记忆/流程），禁止一刀切「收敛到单一工作区」。\",\n \"version_note\": \"示例：对齐用户定稿模板结构（看板表 + 主题分节 + 动作台账 + 可选事实更正）。\",\n \"task_split\": \"复盘会话只做复盘；公众号等成品放独立会话。\"\n },\n \"source\": {\n \"label\": \"agent-collected\",\n \"note\": \"宿主 Agent 采集的会话/用量事实（只读）；分项目须 Glob/ls 核对磁盘后再下结论\"\n },\n \"dashboard\": {\n \"metrics\": [\n {\n \"name\": \"原始 wall-clock 时长\",\n \"value\": \"**≈220h**（在周内新建会话）\",\n \"note\": \"不可用：被跨周/跨边界长会话拉高\"\n },\n {\n \"name\": \"**真实活跃时长**\",\n \"value\": \"**≈39h**\",\n \"note\": \"已核对时间线，剔除跨夜 idle 与跨周只计在周活跃段\"\n },\n {\n \"name\": \"Token 消耗\",\n \"value\": \"**≈270万**\",\n \"note\": \"近四周新高；以 Agent 汇总口径为准\"\n },\n {\n \"name\": \"会话数\",\n \"value\": \"**34 个**新建于周内\",\n \"note\": \"含跨周旧会话重新打开\"\n },\n {\n \"name\": \"主力主题\",\n \"value\": \"社媒 Agent + 视频 + 复盘自动化\",\n \"note\": \"三者占真实投入主要部分\"\n },\n {\n \"name\": \"**本周重大更正**\",\n \"value\": \"原「多工作区分散」**前提不成立**\",\n \"note\": \"须 Glob 核对磁盘后再归因\"\n }\n ],\n \"summary\": \"全周只围绕少数主题；最该记的不是「产出」，而是**复盘前先核对文件系统再下结论**。\",\n \"active_hours\": 39\n },\n \"projects\": [\n {\n \"title\": \"社媒 Agent 主题（≈15.4h / ~20 会话）\",\n \"dimensions\": {\n \"做了什么\": \"规划自动内容发布；安装分析 CLI 并出报告；商业可行性探讨；写操作手册/测试方案\",\n \"DB 记录的工作目录\": \"多个 cwd 别名（须与磁盘核对）\",\n \"磁盘真实情况（已核对）\": \"仅少数真实目录存在；其余 cwd 可能为清理后的临时路径\",\n \"低效条目\": \"【用户】寻找类指令缺交付物；【AI】开放式探索缺收敛\",\n \"提示词怎么改\": \"寻找 X → 在项目里调研并产出对比表+首选 1 个+理由\"\n }\n },\n {\n \"title\": \"视频生成主题（≈13.6h / 10 会话）\",\n \"dimensions\": {\n \"做了什么\": \"分析视频问题；教程生成；发布排查；替代方案探讨\",\n \"DB 记录的工作目录\": \"视频相关 cwd 若干\",\n \"磁盘真实情况（已核对）\": \"仅一个真实视频工作目录\",\n \"低效条目\": \"【用户】分析类缺交付物；【AI】跨夜 idle 缺退出\",\n \"提示词怎么改\": \"分析类输出短结论+避坑清单；生成类加时间盒\"\n }\n }\n ],\n \"other_projects\": [\n {\n \"name\": \"复盘项目(自动化)\",\n \"hours\": \"≈8.2h\",\n \"note\": \"正向：机制落地；单任务过长应加 checkpoint\"\n },\n {\n \"name\": \"测试专家\",\n \"hours\": \"≈1.5h\",\n \"note\": \"干净：目标清晰无漂移\"\n },\n {\n \"name\": \"其他零散\",\n \"hours\": \"≈0.5h\",\n \"note\": \"短任务\"\n }\n ],\n \"charts_data\": {\n \"time_distribution\": [\n { \"label\": \"社媒 Agent 主题\", \"hours\": 15.4 },\n { \"label\": \"视频生成主题\", \"hours\": 13.6 },\n { \"label\": \"复盘项目(自动化)\", \"hours\": 8.2 },\n { \"label\": \"测试专家\", \"hours\": 1.5 },\n { \"label\": \"其他零散\", \"hours\": 0.5 }\n ],\n \"attribution\": [\n { \"label\": \"【用户】指令模糊\", \"count\": 3 },\n { \"label\": \"【AI】缺收敛约束\", \"count\": 3 },\n { \"label\": \"合理分工缺索引\", \"count\": 1 },\n { \"label\": \"重复投入（待核实）\", \"count\": 1 }\n ],\n \"token_trend\": [\n { \"label\": \"前两周\", \"millions\": 0.34 },\n { \"label\": \"上周\", \"millions\": 0.87 },\n { \"label\": \"本周\", \"millions\": 2.74 }\n ]\n },\n \"problems\": [\n {\n \"description\": \"~1/3 指令仍为\\\"动词+无交付物\\\"\",\n \"category\": \"【用户】\",\n \"root_cause\": \"提问习惯（思路）\",\n \"suggestion\": \"反问训练机制持续执行\"\n },\n {\n \"description\": \"探索/生成类任务跨夜 idle\",\n \"category\": \"【AI】\",\n \"root_cause\": \"agent 缺收敛约束（流程）\",\n \"suggestion\": \"可行性判定+试错上限+时间盒\"\n }\n ],\n \"positives\": [\n \"指令干净、含工具+动作+交付物的范本会话\",\n \"主动核对文件系统、推翻错误归因\"\n ],\n \"actions\": {\n \"done\": [\n {\n \"action\": \"会话标识改为可读格式\",\n \"trigger\": \"用户说编号找不到\",\n \"change\": \"改用「标题前30字+创建时间+目录末段」\"\n },\n {\n \"action\": \"新增长会话主动对齐节\",\n \"trigger\": \"用户要求当场拍板\",\n \"change\": \"并入第五节待对齐开放会话\"\n }\n ],\n \"observing\": [\n {\n \"action\": \"反问训练机制上线\",\n \"criteria\": \"下周模糊指令占比是否明显下降？\"\n },\n {\n \"action\": \"复盘先核对文件系统再下结论\",\n \"criteria\": \"下周是否还有仅凭 cwd 臆断目录结构？\"\n }\n ],\n \"pending\": [\n {\n \"suggestion\": \"跨夜 idle 会话继续对齐\",\n \"priority\": \"中\",\n \"deadline\": \"2026-07-26\",\n \"status\": \"待开始\"\n }\n ]\n },\n \"open_sessions\": [\n {\n \"label\": \"「评价 Agent-Skill 识别准确性…」（7/17, 默认文件）\",\n \"span\": \"63h\",\n \"decision\": \"继续推进\",\n \"follow_up\": \"做完自关；AI 追踪到关闭\"\n },\n {\n \"label\": \"TubePilot 探讨会话（教程视频调优, 7/16→7/17）\",\n \"span\": \"12.1h\",\n \"decision\": \"跨夜 idle，计入缺退出机制\",\n \"follow_up\": \"下周继续对齐\"\n }\n ],\n \"automations\": {\n \"narrative\": [\n \"**周度复盘自动化**：周期内触发记录与结果由 Agent 据本平台事实填写；失败/超时需区分调度 off-by-one 与真实故障。\",\n \"**建议**：保持现状；连续断档再评估告警。\"\n ]\n },\n \"corrections_lesson\": \"执行待落实项前，先 Glob 核对文件系统；否则可能把不存在的「分散」当成真问题。\",\n \"corrections\": [\n {\n \"was\": \"主题拆多个工作区\",\n \"claimed\": \"多个 DB cwd\",\n \"actual\": \"磁盘仅少数真实目录存在\"\n }\n ],\n \"corrections_takeaway\": \"复盘归因不能只信会话库 cwd 字符串，必须核对磁盘真实目录再下结论。\"\n}\n\nArchive v1.2.4: 13 files, 24802 bytes\n\nFiles: charts.py (13406b), cli.py (3373b), legacy/analyzer.py (10570b), legacy/README.md (312b), mcp_server.py (3991b), README.md (998b), report.py (15588b), schema/README.md (1652b), schema/review-input.example.json (6932b), skill-card.md (2497b), SKILL.md (2913b), utils.py (506b), _meta.json (132b)\n\nFile v1.2.4:SKILL.md\n\n---\nname: weekly-review\ndescription: >\n 面向任意 AI Agent 自动复盘的 Skill（weekly-review / 复盘 / 提示词优化 / 会话清理）。\n\n 核心能力覆盖：AI 用量与时间看板、提示词复盘与改写建议、高效/低效归因、\n 时间分布与 Token/归因趋势图、开放会话对齐清理、定时自动周报。\n\n 当用户任务涉及周度复盘、用量查看、提示词改进时使用，包括但不限于：\n Token 花了多少、提示词优化改进、一周总结、工作复盘、复盘可视化报告、\n 清理对话、AI 自动复盘。\nversion: 1.2.4\ncategory: 办公效率\nread_when:\n - 提示词改进\n - 本周 AI 用量看板\n - 高低效归因\n - 复盘可视化报告\n - 清理开放对话\n - 定时/自动周报\n - weekly-review / 复盘 / /weekly-review\nmetadata:\n openclaw:\n requires:\n bins:\n - python\n emoji: \"📊\"\n homepage: https://github.com/testman2025/weekly-review-skill\n os:\n - macos\n - linux\n - windows\n---\n\n# AI 用量与提示词复盘助手（weekly-review）\n\n用来做：提示词改进、本周 AI 用量看板、高低效归因、复盘可视化图表、开放会话清理、定时/自动周报。宿主 Agent 采数，本 skill 按六章模板出报告。\n\n## 输出结构（锁定）\n\n1. **一页看板**：`| 指标 | 数值 | 口径说明 |` + **一句话结论** + 可选「辅助图表」PNG 路径 \n2. **分项目分析**：`### 2.x 主题（时长/会话）` + `| 维度 | 内容 |`（做了什么 / DB cwd / 磁盘核对 / 低效 / 提示词）；其他用简表 \n3. **问题清单 + 根因三类归因** + **做得好的（正面范本）** \n4. **本周动作台账 ★**：当场已改 / 观察 / 待落实（或待落实→执行结果） \n5. **待对齐开放会话** \n6. **自动化概览** \n7. **事实更正**（可选，有核对推翻时必写）\n\n**图表约定**：渲染器生成两张定稿样式图——横向「时间投入分布」、双栏「低效归因 + Token 趋势」（SVG）。数据来自 `charts_data`。看板主表仍是 `| 指标 | 数值 | 口径说明 |`。\n\n## 输入\n\n见 `schema/review-input.example.json` 与 `schema/README.md`。\n\n## 操作流程\n\n1. 安装本 skill（见仓库根 README：`npx skills add` / ClawHub / clone）。 \n2. Agent 采集会话/用量，**Glob/ls 核对磁盘**后再归因。 \n3. 填 `review-input`（或对话等价结构）；需要图时先出 PNG 再填 `dashboard.charts`。 \n4. 按六章输出，或 `python -m cli --input review-input.json -o YYYY-MM-DD周度复盘.md`。 \n5. 对开放会话当场拍板。定时场景见仓库 `automations/`。\n\n### CLI\n\n```bash\ncd {SKILL_DIR}\npython -m cli --input schema/review-input.example.json -o 周度复盘.md\n# 图表默认写到报告同目录；关闭：加 --no-charts\n```\n\n### MCP\n\n工具 `run_weekly_review`，参数 `review_input`（对象）。\n\nFile v1.2.4:legacy/README.md\n\n# Legacy（非公开主路径）\n\n此处保留旧版「直接读 SQLite 会话库」分析器，仅供对照或迁移。\n\n**公开能力请使用：**\n\n1. Agent 按 `SKILL.md` 采集本平台数据 \n2. 填入 `schema/review-input.example.json` 结构 \n3. `python -m cli --input review-input.json -o report.md`\n\nFile v1.2.4:README.md\n\n# AI 用量与提示词复盘助手（weekly-review）\n\n面向任意 AI Agent 的自动复盘 Skill。\n\n## 解决什么问题\n\n1. 提升提示词表达能力 \n2. 查看本周 AI 用量 \n3. 总结工作低效与高效 \n4. 复盘可视化图表（时间分布 / 归因 / Token） \n5. 清理/对齐开放对话 \n6. AI 自动/定时周复盘 \n\n## Install\n\n```bash\nnpx skills add testman2025/weekly-review-skill --skill weekly-review\n# or\nopenclaw skills install weekly-review\n```\n\nSee the repo root [README](../../README.md).\n\n## Output shape (locked)\n\n1. Dashboard table (`指标 | 数值 | 口径说明`) + one-liner + optional chart paths \n2. Theme project sections with dimension tables (incl. disk verification) \n3. Problems + three-class root cause + positives \n4. Action ledger \n5. Open sessions \n6. Automations overview \n7. Fact corrections (optional)\n\n## Render\n\n```bash\npython -m cli --input schema/review-input.example.json -o weekly-report.md\n```\n\nSee `SKILL.md` and `schema/`.\n\nFile v1.2.4:schema/README.md\n\n# review-input 约定\n\n对齐定稿模板：`# YYYY-MM-DD 周度复盘（weekly-review skill · 六章结构）`。\n\nAgent 采集事实后填本结构；skill 只渲染结构。**图表**用 `dashboard.charts` 引用已生成的 PNG 路径（辅助图表），不以 Mermaid 为主。\n\n## 图表\n\n渲染默认生成两张**定稿样式 SVG**（标准库）：\n\n1. `chart-时间分布-*.svg` — 横向条形「本周时间投入分布」\n2. `chart-归因Token-*.svg` — 左右双柱「低效条目归因」+「Token 消耗趋势」\n\n数据优先读 `charts_data`：\n\n```json\n\"charts_data\": {\n \"time_distribution\": [{ \"label\": \"社媒 Agent 主题\", \"hours\": 15.4 }],\n \"attribution\": [{ \"label\": \"【用户】指令模糊\", \"count\": 3 }],\n \"token_trend\": [{ \"label\": \"本周\", \"millions\": 2.74 }]\n}\n```\n\n报告用 `![...](path)` 嵌入。`--no-charts` 可关闭。\n\n\n| 字段 | 说明 |\n|------|------|\n| `period` | `start` / `end`（必填） |\n| `meta` | `title` / `methodology` / `version_note` / `task_split` |\n| `source.note` | 数据来源一行说明 |\n| `dashboard.metrics[]` | `{name,value,note}` 看板表 |\n| `dashboard.summary` | 一句话结论 |\n| `dashboard.charts[]` | 辅助图 PNG 相对路径 |\n| `projects[]` | `{title, dimensions{做了什么,…}}` 主题分节 |\n| `other_projects[]` | `{name,hours,note}` |\n| `problems[]` / `positives[]` | 问题表 + 正面范本 |\n| `actions` | `done` / `observing` / `pending` |\n| `open_sessions[]` | 待对齐会话 |\n| `automations.narrative` | 自动化概览（条目列表或长文） |\n| `corrections*` | 可选第七章事实更正 |\n\n完整示例：`review-input.example.json`。\n\nFile v1.2.4:_meta.json\n\n{\n \"ownerId\": \"kn71svn18t9cb73emms0vf78s984tb7p\",\n \"slug\": \"weekly-review\",\n \"version\": \"1.2.4\",\n \"publishedAt\": 1784863773368\n}\n\nFile v1.2.4:skill-card.md\n\n## Description:
\nWeekly Review helps agents turn collected AI usage and session-review facts into weekly review reports with dashboards, prompt-improvement analysis, charts, action ledgers, open-session alignment, and automation summaries.
\n\nThis skill is ready for commercial/non-commercial use.
\n\n## Publisher:
\n[testman2025](https://clawhub.ai/user/testman2025)
\n\n### License/Terms of Use:
\nMIT-0
\n\n\n## Use Case:
\nEmployees, external users, and developers use this skill to review AI work patterns, improve prompts, summarize inefficient sessions, render weekly Markdown reports, and decide whether to clean up open conversations.
\n\n### Deployment Geography for Use:
\nGlobal
\n\n## Known Risks and Mitigations:
\nRisk: The skill may process sensitive AI usage and session-review data.
\nMitigation: Decide in advance which sessions, folders, and usage records the host agent may inspect, and provide only that approved review input.
\nRisk: Generated reports and charts may write sensitive review details to the selected output location.
\nMitigation: Verify the report and chart output paths before running the CLI or MCP workflow.
\nRisk: Scheduled reports or conversation cleanup can affect ongoing workflows if enabled unintentionally.
\nMitigation: Enable scheduled reporting or conversation cleanup only when those actions are explicitly desired.
\n\n\n## Reference(s):
\n- [ClawHub skill page](https://clawhub.ai/testman2025/skills/weekly-review)
\n- [Project homepage](https://github.com/testman2025/weekly-review-skill)
\n- [Input schema](artifact/schema/README.md)
\n- [Example review input](artifact/schema/review-input.example.json)
\n\n\n## Skill Output:
\n**Output Type(s):** [text, markdown, code, shell commands, configuration, guidance]
\n**Output Format:** [Markdown reports, JSON echoes, SVG chart files, and MCP text responses]
\n**Output Parameters:** [1D]
\n**Other Properties Related to Output:** [Can generate weekly-review Markdown from review-input JSON and optional SVG charts unless chart generation is disabled.]
\n\n## Skill Version(s):
\n1.2.4 (source: server release and skill frontmatter)
\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\nFile v1.2.4:schema/review-input.example.json\n\n{\n \"period\": {\n \"start\": \"2026-07-13\",\n \"end\": \"2026-07-19\"\n },\n \"meta\": {\n \"title\": \"2026-07-19 周度复盘（weekly-review skill · 六章结构）\",\n \"methodology\": \"本版按 `weekly-review` skill 六章固定结构组织；归因遵循三类框架（思路/记忆/流程），禁止一刀切「收敛到单一工作区」。\",\n \"version_note\": \"示例：对齐用户定稿模板结构（看板表 + 主题分节 + 动作台账 + 可选事实更正）。\",\n \"task_split\": \"复盘会话只做复盘；公众号等成品放独立会话。\"\n },\n \"source\": {\n \"label\": \"agent-collected\",\n \"note\": \"宿主 Agent 采集的会话/用量事实（只读）；分项目须 Glob/ls 核对磁盘后再下结论\"\n },\n \"dashboard\": {\n \"metrics\": [\n {\n \"name\": \"原始 wall-clock 时长\",\n \"value\": \"**≈220h**（在周内新建会话）\",\n \"note\": \"不可用：被跨周/跨边界长会话拉高\"\n },\n {\n \"name\": \"**真实活跃时长**\",\n \"value\": \"**≈39h**\",\n \"note\": \"已核对时间线，剔除跨夜 idle 与跨周只计在周活跃段\"\n },\n {\n \"name\": \"Token 消耗\",\n \"value\": \"**≈270万**\",\n \"note\": \"近四周新高；以 Agent 汇总口径为准\"\n },\n {\n \"name\": \"会话数\",\n \"value\": \"**34 个**新建于周内\",\n \"note\": \"含跨周旧会话重新打开\"\n },\n {\n \"name\": \"主力主题\",\n \"value\": \"社媒 Agent + 视频 + 复盘自动化\",\n \"note\": \"三者占真实投入主要部分\"\n },\n {\n \"name\": \"**本周重大更正**\",\n \"value\": \"原「多工作区分散」**前提不成立**\",\n \"note\": \"须 Glob 核对磁盘后再归因\"\n }\n ],\n \"summary\": \"全周只围绕少数主题；最该记的不是「产出」，而是**复盘前先核对文件系统再下结论**。\",\n \"active_hours\": 39\n },\n \"projects\": [\n {\n \"title\": \"社媒 Agent 主题（≈15.4h / ~20 会话）\",\n \"dimensions\": {\n \"做了什么\": \"规划自动内容发布；安装分析 CLI 并出报告；商业可行性探讨；写操作手册/测试方案\",\n \"DB 记录的工作目录\": \"多个 cwd 别名（须与磁盘核对）\",\n \"磁盘真实情况（已核对）\": \"仅少数真实目录存在；其余 cwd 可能为清理后的临时路径\",\n \"低效条目\": \"【用户】寻找类指令缺交付物；【AI】开放式探索缺收敛\",\n \"提示词怎么改\": \"寻找 X → 在项目里调研并产出对比表+首选 1 个+理由\"\n }\n },\n {\n \"title\": \"视频生成主题（≈13.6h / 10 会话）\",\n \"dimensions\": {\n \"做了什么\": \"分析视频问题；教程生成；发布排查；替代方案探讨\",\n \"DB 记录的工作目录\": \"视频相关 cwd 若干\",\n \"磁盘真实情况（已核对）\": \"仅一个真实视频工作目录\",\n \"低效条目\": \"【用户】分析类缺交付物；【AI】跨夜 idle 缺退出\",\n \"提示词怎么改\": \"分析类输出短结论+避坑清单；生成类加时间盒\"\n }\n }\n ],\n \"other_projects\": [\n {\n \"name\": \"复盘项目(自动化)\",\n \"hours\": \"≈8.2h\",\n \"note\": \"正向：机制落地；单任务过长应加 checkpoint\"\n },\n {\n \"name\": \"测试专家\",\n \"hours\": \"≈1.5h\",\n \"note\": \"干净：目标清晰无漂移\"\n },\n {\n \"name\": \"其他零散\",\n \"hours\": \"≈0.5h\",\n \"note\": \"短任务\"\n }\n ],\n \"charts_data\": {\n \"time_distribution\": [\n { \"label\": \"社媒 Agent 主题\", \"hours\": 15.4 },\n { \"label\": \"视频生成主题\", \"hours\": 13.6 },\n { \"label\": \"复盘项目(自动化)\", \"hours\": 8.2 },\n { \"label\": \"测试专家\", \"hours\": 1.5 },\n { \"label\": \"其他零散\", \"hours\": 0.5 }\n ],\n \"attribution\": [\n { \"label\": \"【用户】指令模糊\", \"count\": 3 },\n { \"label\": \"【AI】缺收敛约束\", \"count\": 3 },\n { \"label\": \"合理分工缺索引\", \"count\": 1 },\n { \"label\": \"重复投入（待核实）\", \"count\": 1 }\n ],\n \"token_trend\": [\n { \"label\": \"前两周\", \"millions\": 0.34 },\n { \"label\": \"上周\", \"millions\": 0.87 },\n { \"label\": \"本周\", \"millions\": 2.74 }\n ]\n },\n \"problems\": [\n {\n \"description\": \"~1/3 指令仍为\\\"动词+无交付物\\\"\",\n \"category\": \"【用户】\",\n \"root_cause\": \"提问习惯（思路）\",\n \"suggestion\": \"反问训练机制持续执行\"\n },\n {\n \"description\": \"探索/生成类任务跨夜 idle\",\n \"category\": \"【AI】\",\n \"root_cause\": \"agent 缺收敛约束（流程）\",\n \"suggestion\": \"可行性判定+试错上限+时间盒\"\n }\n ],\n \"positives\": [\n \"指令干净、含工具+动作+交付物的范本会话\",\n \"主动核对文件系统、推翻错误归因\"\n ],\n \"actions\": {\n \"done\": [\n {\n \"action\": \"会话标识改为可读格式\",\n \"trigger\": \"用户说编号找不到\",\n \"change\": \"改用「标题前30字+创建时间+目录末段」\"\n },\n {\n \"action\": \"新增长会话主动对齐节\",\n \"trigger\": \"用户要求当场拍板\",\n \"change\": \"并入第五节待对齐开放会话\"\n }\n ],\n \"observing\": [\n {\n \"action\": \"反问训练机制上线\",\n \"criteria\": \"下周模糊指令占比是否明显下降？\"\n },\n {\n \"action\": \"复盘先核对文件系统再下结论\",\n \"criteria\": \"下周是否还有仅凭 cwd 臆断目录结构？\"\n }\n ],\n \"pending\": [\n {\n \"suggestion\": \"跨夜 idle 会话继续对齐\",\n \"priority\": \"中\",\n \"deadline\": \"2026-07-26\",\n \"status\": \"待开始\"\n }\n ]\n },\n \"open_sessions\": [\n {\n \"label\": \"「评价 Agent-Skill 识别准确性…」（7/17, 默认文件）\",\n \"span\": \"63h\",\n \"decision\": \"继续推进\",\n \"follow_up\": \"做完自关；AI 追踪到关闭\"\n },\n {\n \"label\": \"TubePilot 探讨会话（教程视频调优, 7/16→7/17）\",\n \"span\": \"12.1h\",\n \"decision\": \"跨夜 idle，计入缺退出机制\",\n \"follow_up\": \"下周继续对齐\"\n }\n ],\n \"automations\": {\n \"narrative\": [\n \"**周度复盘自动化**：周期内触发记录与结果由 Agent 据本平台事实填写；失败/超时需区分调度 off-by-one 与真实故障。\",\n \"**建议**：保持现状；连续断档再评估告警。\"\n ]\n },\n \"corrections_lesson\": \"执行待落实项前，先 Glob 核对文件系统；否则可能把不存在的「分散」当成真问题。\",\n \"corrections\": [\n {\n \"was\": \"主题拆多个工作区\",\n \"claimed\": \"多个 DB cwd\",\n \"actual\": \"磁盘仅少数真实目录存在\"\n }\n ],\n \"corrections_takeaway\": \"复盘归因不能只信会话库 cwd 字符串，必须核对磁盘真实目录再下结论。\"\n}\n\nArchive v1.2.3: 13 files, 24669 bytes\n\nFiles: charts.py (13406b), cli.py (3373b), legacy/analyzer.py (10570b), legacy/README.md (312b), mcp_server.py (3991b), README.md (702b), report.py (15588b), schema/README.md (1652b), schema/review-input.example.json (6932b), skill-card.md (2352b), SKILL.md (2875b), utils.py (506b), _meta.json (132b)\n\nFile v1.2.3:SKILL.md\n\n---\nname: weekly-review\ndescription: >\n Agent-fed weekly retrospective locked to the six-chapter template\n (dashboard table + one-liner + optional PNG charts, theme projects,\n root-cause triage, action ledger, open sessions, automations, optional\n fact-corrections). Host Agent collects facts; this skill defines structure\n and can render Markdown. Use for weekly-review / 周度复盘 / 本周复盘.\nversion: 1.2.3\ncategory: 办公效率\nread_when:\n - 用户说\"做周度复盘 / 本周复盘 / 跑一下 weekly review\"\n - 用户想清理或对齐跨周、跨夜还开着的会话\n - 用户希望把复盘流程标准化、沉淀成可复用模板\n - 每周固定时间（如周日）触发复盘\nmetadata:\n openclaw:\n requires:\n bins:\n - python\n emoji: \"📊\"\n homepage: https://github.com/testman2025/weekly-review-skill\n os:\n - macos\n - linux\n - windows\n---\n\n# weekly-review（周度复盘 skill）\n\n**固化底座，不固化成品。** 宿主 Agent 采集本平台事实；输出必须贴合定稿六章模板（参考用户已定稿的周度复盘 Markdown）。\n\n## 适用 / 不适用\n\n- **适用**：每周定期复盘；长会话对齐；标准化归因与台账。\n- **不适用**：实时监控；非会话类复盘；替你写公众号成品；替 Agent 读各平台私有存储。\n\n## 输出结构（锁定）\n\n1. **一页看板**：`| 指标 | 数值 | 口径说明 |` + **一句话结论** + 可选「辅助图表」PNG 路径 \n2. **分项目分析**：`### 2.x 主题（时长/会话）` + `| 维度 | 内容 |`（做了什么 / DB cwd / 磁盘核对 / 低效 / 提示词）；其他用简表 \n3. **问题清单 + 根因三类归因** + **做得好的（正面范本）** \n4. **本周动作台账 ★**：当场已改 / 观察 / 待落实（或待落实→执行结果） \n5. **待对齐开放会话** \n6. **自动化概览** \n7. **事实更正**（可选，有核对推翻时必写）\n\n**图表约定**：渲染器生成两张定稿样式图——横向「时间投入分布」、双栏「低效归因 + Token 趋势」（SVG）。数据来自 `charts_data`。看板主表仍是 `| 指标 | 数值 | 口径说明 |`。\n\n## 输入\n\n见 `schema/review-input.example.json` 与 `schema/README.md`。\n\n## 操作流程\n\n1. 安装本 skill。 \n2. Agent 采集会话/用量，**Glob/ls 核对磁盘**后再归因。 \n3. 填 `review-input`（或对话等价结构）；需要图时先出 PNG 再填 `dashboard.charts`。 \n4. 按六章输出，或 `python -m cli --input review-input.json -o YYYY-MM-DD周度复盘.md`。 \n5. 对开放会话当场拍板。\n\n### CLI\n\n```bash\ncd {SKILL_DIR}\npython -m cli --input schema/review-input.example.json -o 周度复盘.md\n# 图表默认写到报告同目录；关闭：加 --no-charts\n```\n\n### MCP\n\n工具 `run_weekly_review`，参数 `review_input`（对象）。\n\nFile v1.2.3:legacy/README.md\n\n# Legacy（非公开主路径）\n\n此处保留旧版「直接读 SQLite 会话库」分析器，仅供对照或迁移。\n\n**公开能力请使用：**\n\n1. Agent 按 `SKILL.md` 采集本平台数据 \n2. 填入 `schema/review-input.example.json` 结构 \n3. `python -m cli --input review-input.json -o report.md`\n\nFile v1.2.3:README.md\n\n# Weekly Review Skill\n\nHost Agent collects facts; this skill locks the **six-chapter Markdown template** used in production retrospectives.\n\n## Output shape (locked)\n\n1. Dashboard table (`指标 | 数值 | 口径说明`) + one-liner + optional PNG chart paths \n2. Theme project sections with dimension tables (incl. disk verification) \n3. Problems + three-class root cause + positives \n4. Action ledger \n5. Open sessions \n6. Automations overview \n7. Fact corrections (optional)\n\nCharts = PNG paths under 辅助图表 — not Mermaid as the primary layout.\n\n## Install / render\n\n```bash\npython -m cli --input schema/review-input.example.json -o weekly-report.md\n```\n\nSee `SKILL.md` and `schema/`.\n\nFile v1.2.3:schema/README.md\n\n# review-input 约定\n\n对齐定稿模板：`# YYYY-MM-DD 周度复盘（weekly-review skill · 六章结构）`。\n\nAgent 采集事实后填本结构；skill 只渲染结构。**图表**用 `dashboard.charts` 引用已生成的 PNG 路径（辅助图表），不以 Mermaid 为主。\n\n## 图表\n\n渲染默认生成两张**定稿样式 SVG**（标准库）：\n\n1. `chart-时间分布-*.svg` — 横向条形「本周时间投入分布」\n2. `chart-归因Token-*.svg` — 左右双柱「低效条目归因」+「Token 消耗趋势」\n\n数据优先读 `charts_data`：\n\n```json\n\"charts_data\": {\n \"time_distribution\": [{ \"label\": \"社媒 Agent 主题\", \"hours\": 15.4 }],\n \"attribution\": [{ \"label\": \"【用户】指令模糊\", \"count\": 3 }],\n \"token_trend\": [{ \"label\": \"本周\", \"millions\": 2.74 }]\n}\n```\n\n报告用 `![...](path)` 嵌入。`--no-charts` 可关闭。\n\n\n| 字段 | 说明 |\n|------|------|\n| `period` | `start` / `end`（必填） |\n| `meta` | `title` / `methodology` / `version_note` / `task_split` |\n| `source.note` | 数据来源一行说明 |\n| `dashboard.metrics[]` | `{name,value,note}` 看板表 |\n| `dashboard.summary` | 一句话结论 |\n| `dashboard.charts[]` | 辅助图 PNG 相对路径 |\n| `projects[]` | `{title, dimensions{做了什么,…}}` 主题分节 |\n| `other_projects[]` | `{name,hours,note}` |\n| `problems[]` / `positives[]` | 问题表 + 正面范本 |\n| `actions` | `done` / `observing` / `pending` |\n| `open_sessions[]` | 待对齐会话 |\n| `automations.narrative` | 自动化概览（条目列表或长文） |\n| `corrections*` | 可选第七章事实更正 |\n\n完整示例：`review-input.example.json`。\n\nFile v1.2.3:_meta.json\n\n{\n \"ownerId\": \"kn71svn18t9cb73emms0vf78s984tb7p\",\n \"slug\": \"weekly-review\",\n \"version\": \"1.2.3\",\n \"publishedAt\": 1784611687772\n}\n\nFile v1.2.3:skill-card.md\n\n## Description:
\nWeekly Review helps an agent turn collected session and usage facts into a fixed six-chapter retrospective with dashboard metrics, project analysis, root-cause triage, action ledgers, open sessions, automations, and optional chart assets.
\n\nThis skill is ready for commercial/non-commercial use.
\n\n## Publisher:
\n[testman2025](https://clawhub.ai/user/testman2025)
\n\n### License/Terms of Use:
\nMIT-0
\n\n\n## Use Case:
\nDevelopers and agent users use this skill to standardize weekly retrospectives from already collected session, usage, project, action, and automation facts. It can render the supplied review-input JSON into Markdown and optional chart files.
\n\n### Deployment Geography for Use:
\nGlobal
\n\n## Known Risks and Mitigations:
\nRisk: The package includes legacy code capable of reading a local AI-session SQLite database when explicitly invoked or when database environment variables are set.
\nMitigation: Use the documented review-input JSON, CLI renderer, or MCP renderer paths by default, and invoke the legacy analyzer only after intentionally approving the database path and the session data it will process.
\n\n\n## Reference(s):
\n- [ClawHub skill page](https://clawhub.ai/testman2025/skills/weekly-review)
\n- [Project homepage](https://github.com/testman2025/weekly-review-skill)
\n- [Skill README](artifact/README.md)
\n- [Review input schema](artifact/schema/README.md)
\n- [Review input example](artifact/schema/review-input.example.json)
\n\n\n## Skill Output:
\n**Output Type(s):** [text, markdown, code, shell commands, configuration, guidance]
\n**Output Format:** [Markdown report, JSON echo mode, MCP text response, and optional SVG chart files]
\n**Output Parameters:** [1D]
\n**Other Properties Related to Output:** [Requires a review-input JSON object supplied by the host agent; optional chart rendering writes two SVG assets unless disabled.]
\n\n## Skill Version(s):
\n1.2.3 (source: server release evidence and SKILL.md frontmatter)
\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\nFile v1.2.3:schema/review-input.example.json\n\n{\n \"period\": {\n \"start\": \"2026-07-13\",\n \"end\": \"2026-07-19\"\n },\n \"meta\": {\n \"title\": \"2026-07-19 周度复盘（weekly-review skill · 六章结构）\",\n \"methodology\": \"本版按 `weekly-review` skill 六章固定结构组织；归因遵循三类框架（思路/记忆/流程），禁止一刀切「收敛到单一工作区」。\",\n \"version_note\": \"示例：对齐用户定稿模板结构（看板表 + 主题分节 + 动作台账 + 可选事实更正）。\",\n \"task_split\": \"复盘会话只做复盘；公众号等成品放独立会话。\"\n },\n \"source\": {\n \"label\": \"agent-collected\",\n \"note\": \"宿主 Agent 采集的会话/用量事实（只读）；分项目须 Glob/ls 核对磁盘后再下结论\"\n },\n \"dashboard\": {\n \"metrics\": [\n {\n \"name\": \"原始 wall-clock 时长\",\n \"value\": \"**≈220h**（在周内新建会话）\",\n \"note\": \"不可用：被跨周/跨边界长会话拉高\"\n },\n {\n \"name\": \"**真实活跃时长**\",\n \"value\": \"**≈39h**\",\n \"note\": \"已核对时间线，剔除跨夜 idle 与跨周只计在周活跃段\"\n },\n {\n \"name\": \"Token 消耗\",\n \"value\": \"**≈270万**\",\n \"note\": \"近四周新高；以 Agent 汇总口径为准\"\n },\n {\n \"name\": \"会话数\",\n \"value\": \"**34 个**新建于周内\",\n \"note\": \"含跨周旧会话重新打开\"\n },\n {\n \"name\": \"主力主题\",\n \"value\": \"社媒 Agent + 视频 + 复盘自动化\",\n \"note\": \"三者占真实投入主要部分\"\n },\n {\n \"name\": \"**本周重大更正**\",\n \"value\": \"原「多工作区分散」**前提不成立**\",\n \"note\": \"须 Glob 核对磁盘后再归因\"\n }\n ],\n \"summary\": \"全周只围绕少数主题；最该记的不是「产出」，而是**复盘前先核对文件系统再下结论**。\",\n \"active_hours\": 39\n },\n \"projects\": [\n {\n \"title\": \"社媒 Agent 主题（≈15.4h / ~20 会话）\",\n \"dimensions\": {\n \"做了什么\": \"规划自动内容发布；安装分析 CLI 并出报告；商业可行性探讨；写操作手册/测试方案\",\n \"DB 记录的工作目录\": \"多个 cwd 别名（须与磁盘核对）\",\n \"磁盘真实情况（已核对）\": \"仅少数真实目录存在；其余 cwd 可能为清理后的临时路径\",\n \"低效条目\": \"【用户】寻找类指令缺交付物；【AI】开放式探索缺收敛\",\n \"提示词怎么改\": \"寻找 X → 在项目里调研并产出对比表+首选 1 个+理由\"\n }\n },\n {\n \"title\": \"视频生成主题（≈13.6h / 10 会话）\",\n \"dimensions\": {\n \"做了什么\": \"分析视频问题；教程生成；发布排查；替代方案探讨\",\n \"DB 记录的工作目录\": \"视频相关 cwd 若干\",\n \"磁盘真实情况（已核对）\": \"仅一个真实视频工作目录\",\n \"低效条目\": \"【用户】分析类缺交付物；【AI】跨夜 idle 缺退出\",\n \"提示词怎么改\": \"分析类输出短结论+避坑清单；生成类加时间盒\"\n }\n }\n ],\n \"other_projects\": [\n {\n \"name\": \"复盘项目(自动化)\",\n \"hours\": \"≈8.2h\",\n \"note\": \"正向：机制落地；单任务过长应加 checkpoint\"\n },\n {\n \"name\": \"测试专家\",\n \"hours\": \"≈1.5h\",\n \"note\": \"干净：目标清晰无漂移\"\n },\n {\n \"name\": \"其他零散\",\n \"hours\": \"≈0.5h\",\n \"note\": \"短任务\"\n }\n ],\n \"charts_data\": {\n \"time_distribution\": [\n { \"label\": \"社媒 Agent 主题\", \"hours\": 15.4 },\n { \"label\": \"视频生成主题\", \"hours\": 13.6 },\n { \"label\": \"复盘项目(自动化)\", \"hours\": 8.2 },\n { \"label\": \"测试专家\", \"hours\": 1.5 },\n { \"label\": \"其他零散\", \"hours\": 0.5 }\n ],\n \"attribution\": [\n { \"label\": \"【用户】指令模糊\", \"count\": 3 },\n { \"label\": \"【AI】缺收敛约束\", \"count\": 3 },\n { \"label\": \"合理分工缺索引\", \"count\": 1 },\n { \"label\": \"重复投入（待核实）\", \"count\": 1 }\n ],\n \"token_trend\": [\n { \"label\": \"前两周\", \"millions\": 0.34 },\n { \"label\": \"上周\", \"millions\": 0.87 },\n { \"label\": \"本周\", \"millions\": 2.74 }\n ]\n },\n \"problems\": [\n {\n \"description\": \"~1/3 指令仍为\\\"动词+无交付物\\\"\",\n \"category\": \"【用户】\",\n \"root_cause\": \"提问习惯（思路）\",\n \"suggestion\": \"反问训练机制持续执行\"\n },\n {\n \"description\": \"探索/生成类任务跨夜 idle\",\n \"category\": \"【AI】\",\n \"root_cause\": \"agent 缺收敛约束（流程）\",\n \"suggestion\": \"可行性判定+试错上限+时间盒\"\n }\n ],\n \"positives\": [\n \"指令干净、含工具+动作+交付物的范本会话\",\n \"主动核对文件系统、推翻错误归因\"\n ],\n \"actions\": {\n \"done\": [\n {\n \"action\": \"会话标识改为可读格式\",\n \"trigger\": \"用户说编号找不到\",\n \"change\": \"改用「标题前30字+创建时间+目录末段」\"\n },\n {\n \"action\": \"新增长会话主动对齐节\",\n \"trigger\": \"用户要求当场拍板\",\n \"change\": \"并入第五节待对齐开放会话\"\n }\n ],\n \"observing\": [\n {\n \"action\": \"反问训练机制上线\",\n \"criteria\": \"下周模糊指令占比是否明显下降？\"\n },\n {\n \"action\": \"复盘先核对文件系统再下结论\",\n \"criteria\": \"下周是否还有仅凭 cwd 臆断目录结构？\"\n }\n ],\n \"pending\": [\n {\n \"suggestion\": \"跨夜 idle 会话继续对齐\",\n \"priority\": \"中\",\n \"deadline\": \"2026-07-26\",\n \"status\": \"待开始\"\n }\n ]\n },\n \"open_sessions\": [\n {\n \"label\": \"「评价 Agent-Skill 识别准确性…」（7/17, 默认文件）\",\n \"span\": \"63h\",\n \"decision\": \"继续推进\",\n \"follow_up\": \"做完自关；AI 追踪到关闭\"\n },\n {\n \"label\": \"TubePilot 探讨会话（教程视频调优, 7/16→7/17）\",\n \"span\": \"12.1h\",\n \"decision\": \"跨夜 idle，计入缺退出机制\",\n \"follow_up\": \"下周继续对齐\"\n }\n ],\n \"automations\": {\n \"narrative\": [\n \"**周度复盘自动化**：周期内触发记录与结果由 Agent 据本平台事实填写；失败/超时需区分调度 off-by-one 与真实故障。\",\n \"**建议**：保持现状；连续断档再评估告警。\"\n ]\n },\n \"corrections_lesson\": \"执行待落实项前，先 Glob 核对文件系统；否则可能把不存在的「分散」当成真问题。\",\n \"corrections\": [\n {\n \"was\": \"主题拆多个工作区\",\n \"claimed\": \"多个 DB cwd\",\n \"actual\": \"磁盘仅少数真实目录存在\"\n }\n ],\n \"corrections_takeaway\": \"复盘归因不能只信会话库 cwd 字符串，必须核对磁盘真实目录再下结论。\"\n}\n\nArchive v1.2.1: 12 files, 17060 bytes\n\nFiles: cli.py (1910b), legacy/analyzer.py (10570b), legacy/README.md (312b), mcp_server.py (3991b), README.md (1236b), report.py (9407b), schema/README.md (1044b), schema/review-input.example.json (2157b), skill-card.md (2814b), SKILL.md (4211b), utils.py (506b), _meta.json (132b)\n\nFile v1.2.1:SKILL.md\n\n---\nname: weekly-review\ndescription: >\n Agent-fed weekly retrospective: the host Agent collects session facts from\n its own platform; this skill defines the review structure (dashboard,\n per-project, root-cause triage, action ledger, open sessions) and can render\n Markdown from a standard review-input JSON. Use for weekly-review /\n 周度复盘 / 本周复盘 / 一周总结.\nversion: 1.2.1\ncategory: 办公效率\nread_when:\n - 用户说\"做周度复盘 / 本周复盘 / 跑一下 weekly review\"\n - 用户想清理或对齐跨周、跨夜还开着的会话\n - 用户希望把复盘流程标准化、沉淀成可复用模板\n - 每周固定时间（如周日）触发复盘\nmetadata:\n openclaw:\n requires:\n bins:\n - python\n emoji: \"📊\"\n homepage: https://github.com/testman2025/weekly-review-skill\n os:\n - macos\n - linux\n - windows\n---\n\n# weekly-review（周度复盘 skill）\n\n跨平台周度复盘底座：**各平台 Agent 负责读本机会话/用量并整理事实**；本 skill 负责「复盘看什么、怎么归因、怎么落台账」，以及可选的 Markdown 渲染。\n\n## 适用场景\n\n- **每周定期复盘**，需要客观数据（活跃时长、会话数、用量、跨周/跨夜会话）支撑。\n- **会话 / 工作区越开越多**，需要清理或对齐仍处于 idle 的长会话。\n- **复盘流程标准化**：把「该看什么、怎么归因」沉淀为可复用模板。\n\n## 不适用场景\n\n- **实时会话监控 / 告警**：离线周度分析，不常驻推送。\n- **非会话类复盘**：代码审查质量、财务报表等不在范围内。\n- **替你写公众号/周报成品**：只产出结构化底座，正文由你写。\n- **替 Agent 读各平台私有存储**：本 skill **不**内置各产品数据库适配器；采集由宿主 Agent 完成。\n\n## 设计原则\n\n**固化底座，不固化成品；Agent 采数，Skill 定结构。**\n\n| 职责 | 谁做 |\n|------|------|\n| 读本平台会话 / 用量 / 历史 | **宿主 Agent**（用自己的工具与权限） |\n| 章节结构、根因三类、动作台账 | **本 skill** |\n| 把标准 JSON 打成 Markdown | **本 skill**（可选 CLI / MCP） |\n\n## 输入\n\n标准事实包 `review-input`（对话中填齐亦可），字段见 `schema/README.md` 与 `schema/review-input.example.json`：\n\n- `period`：起止日期（必填）\n- `dashboard`：看板指标\n- `projects[]`：分项目\n- `problems[]` / `actions`：问题与动作台账（常需人工补）\n- `open_sessions[]`：待对齐开放会话\n- `automations`：可选\n\n## 输出\n\n固定六章 Markdown：一页看板、分项目分析、问题+根因三类归因、动作台账、待对齐开放会话、自动化概览。\n\n## 其他约束\n\n- 归因三类：思路 / 记忆 / 流程；禁止一刀切「收敛到单一工作区」。\n- 长会话须先读内容/标题再处置，不能只看时长。\n- 跨周阈值建议 >48h；跨夜 idle 建议次日 06:00 后仍活跃/未关。\n- 可选渲染器仅依赖 Python ≥ 3.10 标准库；无网络、无密钥。\n\n## 如何使用 / 操作流程\n\n1. **安装**本 skill 到所用 Agent 的 skills 目录（或 ClawHub：`openclaw skills install weekly-review`）。\n2. **Agent 采集**：用本平台能力汇总本周会话、用量、项目分布、长会话/跨夜列表（不要假设统一 SQLite）。\n3. **填结构**：写入 `review-input` JSON，或在对话中按同名字段组织；问题与改进建议与用户对齐后写入 `problems` / `actions`。\n4. **出报告**（二选一）：\n - 对话中直接按六章输出 Markdown；或\n - `python -m cli --input review-input.json -o 周度复盘.md`\n5. **拍板**：对 `open_sessions` 逐条确认关闭 / 续作 / 归档。\n\n### CLI 渲染（可选）\n\n```bash\ncd {SKILL_DIR}\npython -m cli --input schema/review-input.example.json -o 周度复盘.md\n```\n\n### MCP（可选）\n\n工具 `run_weekly_review`，参数 `review_input`（对象）。**不接受 db_path**；采集由 Agent 完成后再调用。\n\n```json\n{\n \"mcpServers\": {\n \"weekly-review\": {\n \"command\": \"python\",\n \"args\": [\"-m\", \"mcp_server\"],\n \"env\": { \"PYTHONPATH\": \"{SKILL_DIR}\" }\n }\n }\n}\n```\n\nFile v1.2.1:legacy/README.md\n\n# Legacy（非公开主路径）\n\n此处保留旧版「直接读 SQLite 会话库」分析器，仅供对照或迁移。\n\n**公开能力请使用：**\n\n1. Agent 按 `SKILL.md` 采集本平台数据 \n2. 填入 `schema/review-input.example.json` 结构 \n3. `python -m cli --input review-input.json -o report.md`\n\nFile v1.2.1:README.md\n\n# Weekly Review Skill\n\n**Agent collects facts; this skill defines the retrospective structure** (and optionally renders Markdown).\n\nInstall on any Agent that supports Skills / MCP. The host Agent reads its own platform’s session history; this skill does **not** ship per-vendor DB adapters.\n\n## What it does\n\n- Defines a fixed weekly review: dashboard, per-project, root-cause triage (approach / memory / process), action ledger, open sessions, automations.\n- Accepts a standard `review-input` JSON (see `schema/`).\n- Optional CLI / MCP renderer (Python stdlib only).\n\n## Safety\n\n- No network calls in the renderer.\n- No reading of local session databases by default.\n- No secrets required.\n\n## Install\n\n```bash\nopenclaw skills install weekly-review\n# or\nclawhub install weekly-review\n```\n\n## Usage\n\n1. Agent gathers this week’s session facts from **its own platform**.\n2. Fill `review-input` (see `schema/review-input.example.json`).\n3. Output the six sections in chat, or render:\n\n```bash\ncd \npython -m cli --input schema/review-input.example.json -o weekly-report.md\n```\n\n### MCP\n\nTool `run_weekly_review` requires `review_input` object (not `db_path`).\n\n## License\n\n- Repo source: MIT.\n- ClawHub distribution: MIT-0.\n\nFile v1.2.1:schema/README.md\n\n# review-input 约定\n\nAgent 采集本平台会话事实后，填入本结构；skill 只做复盘流程与可选 Markdown 渲染。\n\n## 字段\n\n| 字段 | 必填 | 说明 |\n|------|------|------|\n| `period.start` / `period.end` | 是 | `YYYY-MM-DD` |\n| `source.label` | 否 | 数据来源标签，如 `agent-collected` |\n| `source.note` | 否 | 补充说明 |\n| `dashboard` | 建议 | 一页看板指标（小时、credits、会话数等） |\n| `projects[]` | 建议 | 分项目：`name`, `sessions`, `hours`, `credits`, `notes` |\n| `problems[]` | 否 | 问题与根因（思路/记忆/流程） |\n| `actions` | 否 | `done` / `observing` / `pending` |\n| `open_sessions[]` | 否 | 跨周 / 跨夜待对齐 |\n| `automations` | 否 | 自动化运行概览 |\n\n完整示例见同目录 `review-input.example.json`。\n\n## 谁负责什么\n\n- **Agent**：用本平台能力读取会话/用量，整理成上述 JSON（或在对话中等价填齐）。\n- **Skill**：规定章节与归因框架；可选 `python -m cli --input …` 渲染 Markdown。\n\nFile v1.2.1:_meta.json\n\n{\n \"ownerId\": \"kn71svn18t9cb73emms0vf78s984tb7p\",\n \"slug\": \"weekly-review\",\n \"version\": \"1.2.1\",\n \"publishedAt\": 1784603331364\n}\n\nFile v1.2.1:skill-card.md\n\n## Description:
\nAgent-fed weekly retrospective: the host Agent collects session facts from its own platform; this skill defines the review structure (dashboard, per-project, root-cause triage, action ledger, open sessions) and can render Markdown from a standard review-input JSON.
\n\nThis skill is ready for commercial/non-commercial use.
\n\n## Publisher:
\n[testman2025](https://clawhub.ai/user/testman2025)
\n\n### License/Terms of Use:
\nMIT-0
\n\n\n## Use Case:
\nDevelopers and agent users use this skill to turn agent-collected weekly session, usage, project, problem, action, and open-session facts into a structured retrospective. It supports chat-based review output and optional local CLI or MCP Markdown rendering from a standard review-input JSON object.
\n\n### Deployment Geography for Use:
\nGlobal
\n\n## Known Risks and Mitigations:
\nRisk: Weekly review inputs can contain sensitive session history, usage patterns, project names, and open work context.
\nMitigation: Have the host agent collect only the facts needed for the retrospective and review the review-input content before rendering or sharing the report.
\nRisk: The legacy SQLite analyzer can read a specified local session database if a user intentionally runs it.
\nMitigation: Use the public Agent-collected review-input workflow by default, and invoke the legacy analyzer only with an explicit database path or environment variable when that local read is intended.
\nRisk: Open sessions may be closed, archived, or continued based on incomplete context if judged only by age.
\nMitigation: Review each long-running session's title and content before deciding whether to close, migrate, continue, or archive it.
\n\n\n## Reference(s):
\n- [ClawHub skill page](https://clawhub.ai/testman2025/skills/weekly-review)
\n- [Project homepage](https://github.com/testman2025/weekly-review-skill)
\n- [review-input schema](schema/README.md)
\n- [review-input example](schema/review-input.example.json)
\n\n\n## Skill Output:
\n**Output Type(s):** [text, markdown, json, guidance]
\n**Output Format:** [Markdown report or JSON text from a standard review-input object]
\n**Output Parameters:** [1D]
\n**Other Properties Related to Output:** [Six-section weekly retrospective covering dashboard metrics, project analysis, root-cause triage, action ledger, open sessions, and automation overview.]
\n\n## Skill Version(s):
\n1.2.1 (source: frontmatter and 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\nFile v1.2.1:schema/review-input.example.json\n\n{\n \"period\": {\n \"start\": \"2026-07-13\",\n \"end\": \"2026-07-19\"\n },\n \"source\": {\n \"label\": \"agent-collected\",\n \"note\": \"由运行本 skill 的 Agent 从本平台会话/用量中整理，非 skill 直接读库\"\n },\n \"dashboard\": {\n \"wall_clock_hours\": 42.5,\n \"active_hours\": 18.2,\n \"credits\": 12.4,\n \"session_count\": 27,\n \"avg_session_hours\": 0.67,\n \"long_session_count\": 2,\n \"overnight_idle_count\": 1,\n \"top_models\": [\n { \"name\": \"composer\", \"count\": 15 },\n { \"name\": \"gpt-5\", \"count\": 8 }\n ]\n },\n \"projects\": [\n {\n \"name\": \"weekly-review-skill\",\n \"sessions\": 8,\n \"hours\": 6.5,\n \"credits\": 3.2,\n \"notes\": \"完成 ClawHub 发布与跨平台文案\"\n },\n {\n \"name\": \"TubePilot\",\n \"sessions\": 5,\n \"hours\": 4.1,\n \"credits\": 2.0,\n \"notes\": \"\"\n }\n ],\n \"problems\": [\n {\n \"description\": \"约 1/3 指令仍为动词+无交付物\",\n \"category\": \"【用户】\",\n \"root_cause\": \"提问习惯\",\n \"suggestion\": \"反问训练机制已写入记忆，持续执行\"\n }\n ],\n \"actions\": {\n \"done\": [\n {\n \"action\": \"长会话处置改为先读内容再判断\",\n \"trigger\": \"误判后纠正\",\n \"change\": \"只看时长会错杀合理分工\"\n }\n ],\n \"observing\": [\n {\n \"action\": \"反问训练机制上线\",\n \"criteria\": \"下周模糊指令占比是否下降\"\n }\n ],\n \"pending\": [\n {\n \"suggestion\": \"跨夜 idle 会话下周继续对齐\",\n \"priority\": \"中\",\n \"deadline\": \"2026-07-26\",\n \"status\": \"待开始\"\n }\n ]\n },\n \"open_sessions\": [\n {\n \"title\": \"探索评估 skill 质量的工具\",\n \"created\": \"07/14\",\n \"last_active\": \"07/16\",\n \"cwd_tail\": \"eval-tools\",\n \"kind\": \"cross_week\",\n \"decision\": \"待拍板\",\n \"follow_up\": \"关闭/迁移/继续\"\n }\n ],\n \"automations\": {\n \"runs_total\": 4,\n \"runs_success\": 3,\n \"items\": [\n {\n \"title\": \"周日复盘提醒\",\n \"status\": \"completed\",\n \"result\": \"成功\",\n \"source\": \"—\"\n }\n ]\n }\n}\n\nArchive v1.2.0: 12 files, 16813 bytes\n\nFiles: cli.py (1910b), legacy/analyzer.py (10570b), legacy/README.md (312b), mcp_server.py (3991b), README.md (1236b), report.py (9407b), schema/README.md (1044b), schema/review-input.example.json (2157b), skill-card.md (2106b), SKILL.md (4211b), utils.py (506b), _meta.json (132b)\n\nFile v1.2.0:SKILL.md\n\n---\nname: weekly-review\ndescription: >\n Agent-fed weekly retrospective: the host Agent collects session facts from\n its own platform; this skill defines the review structure (dashboard,\n per-project, root-cause triage, action ledger, open sessions) and can render\n Markdown from a standard review-input JSON. Use for weekly-review /\n 周度复盘 / 本周复盘 / 一周总结.\nversion: 1.2.0\ncategory: 办公效率\nread_when:\n - 用户说\"做周度复盘 / 本周复盘 / 跑一下 weekly review\"\n - 用户想清理或对齐跨周、跨夜还开着的会话\n - 用户希望把复盘流程标准化、沉淀成可复用模板\n - 每周固定时间（如周日）触发复盘\nmetadata:\n openclaw:\n requires:\n bins:\n - python\n emoji: \"📊\"\n homepage: https://github.com/testman2025/weekly-review-skill\n os:\n - macos\n - linux\n - windows\n---\n\n# weekly-review（周度复盘 skill）\n\n跨平台周度复盘底座：**各平台 Agent 负责读本机会话/用量并整理事实**；本 skill 负责「复盘看什么、怎么归因、怎么落台账」，以及可选的 Markdown 渲染。\n\n## 适用场景\n\n- **每周定期复盘**，需要客观数据（活跃时长、会话数、用量、跨周/跨夜会话）支撑。\n- **会话 / 工作区越开越多**，需要清理或对齐仍处于 idle 的长会话。\n- **复盘流程标准化**：把「该看什么、怎么归因」沉淀为可复用模板。\n\n## 不适用场景\n\n- **实时会话监控 / 告警**：离线周度分析，不常驻推送。\n- **非会话类复盘**：代码审查质量、财务报表等不在范围内。\n- **替你写公众号/周报成品**：只产出结构化底座，正文由你写。\n- **替 Agent 读各平台私有存储**：本 skill **不**内置各产品数据库适配器；采集由宿主 Agent 完成。\n\n## 设计原则\n\n**固化底座，不固化成品；Agent 采数，Skill 定结构。**\n\n| 职责 | 谁做 |\n|------|------|\n| 读本平台会话 / 用量 / 历史 | **宿主 Agent**（用自己的工具与权限） |\n| 章节结构、根因三类、动作台账 | **本 skill** |\n| 把标准 JSON 打成 Markdown | **本 skill**（可选 CLI / MCP） |\n\n## 输入\n\n标准事实包 `review-input`（对话中填齐亦可），字段见 `schema/README.md` 与 `schema/review-input.example.json`：\n\n- `period`：起止日期（必填）\n- `dashboard`：看板指标\n- `projects[]`：分项目\n- `problems[]` / `actions`：问题与动作台账（常需人工补）\n- `open_sessions[]`：待对齐开放会话\n- `automations`：可选\n\n## 输出\n\n固定六章 Markdown：一页看板、分项目分析、问题+根因三类归因、动作台账、待对齐开放会话、自动化概览。\n\n## 其他约束\n\n- 归因三类：思路 / 记忆 / 流程；禁止一刀切「收敛到单一工作区」。\n- 长会话须先读内容/标题再处置，不能只看时长。\n- 跨周阈值建议 >48h；跨夜 idle 建议次日 06:00 后仍活跃/未关。\n- 可选渲染器仅依赖 Python ≥ 3.10 标准库；无网络、无密钥。\n\n## 如何使用 / 操作流程\n\n1. **安装**本 skill 到所用 Agent 的 skills 目录（或 ClawHub：`openclaw skills install weekly-review`）。\n2. **Agent 采集**：用本平台能力汇总本周会话、用量、项目分布、长会话/跨夜列表（不要假设统一 SQLite）。\n3. **填结构**：写入 `review-input` JSON，或在对话中按同名字段组织；问题与改进建议与用户对齐后写入 `problems` / `actions`。\n4. **出报告**（二选一）：\n - 对话中直接按六章输出 Markdown；或\n - `python -m cli --input review-input.json -o 周度复盘.md`\n5. **拍板**：对 `open_sessions` 逐条确认关闭 / 续作 / 归档。\n\n### CLI 渲染（可选）\n\n```bash\ncd {SKILL_DIR}\npython -m cli --input schema/review-input.example.json -o 周度复盘.md\n```\n\n### MCP（可选）\n\n工具 `run_weekly_review`，参数 `review_input`（对象）。**不接受 db_path**；采集由 Agent 完成后再调用。\n\n```json\n{\n \"mcpServers\": {\n \"weekly-review\": {\n \"command\": \"python\",\n \"args\": [\"-m\", \"mcp_server\"],\n \"env\": { \"PYTHONPATH\": \"{SKILL_DIR}\" }\n }\n }\n}\n```\n\nFile v1.2.0:legacy/README.md\n\n# Legacy（非公开主路径）\n\n此处保留旧版「直接读 SQLite 会话库」分析器，仅供对照或迁移。\n\n**公开能力请使用：**\n\n1. Agent 按 `SKILL.md` 采集本平台数据 \n2. 填入 `schema/review-input.example.json` 结构 \n3. `python -m cli --input review-input.json -o report.md`\n\nFile v1.2.0:README.md\n\n# Weekly Review Skill\n\n**Agent collects facts; this skill defines the retrospective structure** (and optionally renders Markdown).\n\nInstall on any Agent that supports Skills / MCP. The host Agent reads its own platform’s session history; this skill does **not** ship per-vendor DB adapters.\n\n## What it does\n\n- Defines a fixed weekly review: dashboard, per-project, root-cause triage (approach / memory / process), action ledger, open sessions, automations.\n- Accepts a standard `review-input` JSON (see `schema/`).\n- Optional CLI / MCP renderer (Python stdlib only).\n\n## Safety\n\n- No network calls in the renderer.\n- No reading of local session databases by default.\n- No secrets required.\n\n## Install\n\n```bash\nopenclaw skills install weekly-review\n# or\nclawhub install weekly-review\n```\n\n## Usage\n\n1. Agent gathers this week’s session facts from **its own platform**.\n2. Fill `review-input` (see `schema/review-input.example.json`).\n3. Output the six sections in chat, or render:\n\n```bash\ncd \npython -m cli --input schema/review-input.example.json -o weekly-report.md\n```\n\n### MCP\n\nTool `run_weekly_review` requires `review_input` object (not `db_path`).\n\n## License\n\n- Repo source: MIT.\n- ClawHub distribution: MIT-0.\n\nFile v1.2.0:schema/README.md\n\n# review-input 约定\n\nAgent 采集本平台会话事实后，填入本结构；skill 只做复盘流程与可选 Markdown 渲染。\n\n## 字段\n\n| 字段 | 必填 | 说明 |\n|------|------|------|\n| `period.start` / `period.end` | 是 | `YYYY-MM-DD` |\n| `source.label` | 否 | 数据来源标签，如 `agent-collected` |\n| `source.note` | 否 | 补充说明 |\n| `dashboard` | 建议 | 一页看板指标（小时、credits、会话数等） |\n| `projects[]` | 建议 | 分项目：`name`, `sessions`, `hours`, `credits`, `notes` |\n| `problems[]` | 否 | 问题与根因（思路/记忆/流程） |\n| `actions` | 否 | `done` / `observing` / `pending` |\n| `open_sessions[]` | 否 | 跨周 / 跨夜待对齐 |\n| `automations` | 否 | 自动化运行概览 |\n\n完整示例见同目录 `review-input.example.json`。\n\n## 谁负责什么\n\n- **Agent**：用本平台能力读取会话/用量，整理成上述 JSON（或在对话中等价填齐）。\n- **Skill**：规定章节与归因框架；可选 `python -m cli --input …` 渲染 Markdown。\n\nFile v1.2.0:_meta.json\n\n{\n \"ownerId\": \"kn71svn18t9cb73emms0vf78s984tb7p\",\n \"slug\": \"weekly-review\",\n \"version\": \"1.2.0\",\n \"publishedAt\": 1784601259113\n}\n\nFile v1.2.0:skill-card.md\n\n## Description:
\nAgent-fed weekly retrospective: the host Agent collects session facts from its own platform; this skill defines the review structure and can render Markdown from a standard review-input JSON.
\n\nThis skill is ready for commercial/non-commercial use.
\n\n## Publisher:
\n[testman2025](https://clawhub.ai/user/testman2025)
\n\n### License/Terms of Use:
\nMIT-0
\n\n\n## Use Case:
\nDevelopers and agent users use this skill to standardize weekly retrospectives from agent-collected session facts, including dashboards, per-project analysis, root-cause triage, action ledgers, open sessions, and automation summaries.
\n\n### Deployment Geography for Use:
\nGlobal
\n\n## Known Risks and Mitigations:
\nRisk: The bundled legacy analyzer can read local session and automation data from a SQLite database.
\nMitigation: Use the normal CLI or MCP renderer with explicit review-input JSON, and do not invoke or expose artifact/legacy/analyzer.py unless local database access is intentional.
\n\n\n## Reference(s):
\n- [ClawHub Skill Page](https://clawhub.ai/testman2025/skills/weekly-review)
\n- [Project Homepage](https://github.com/testman2025/weekly-review-skill)
\n- [review-input schema](artifact/schema/README.md)
\n- [review-input example](artifact/schema/review-input.example.json)
\n\n\n## Skill Output:
\n**Output Type(s):** [text, markdown, json, shell commands, configuration, guidance]
\n**Output Format:** [Markdown reports or validated JSON, with optional CLI and MCP usage guidance.]
\n**Output Parameters:** [1D]
\n**Other Properties Related to Output:** [The normal renderer is local-only and expects an explicit review-input JSON object.]
\n\n## Skill Version(s):
\n1.2.0 (source: SKILL.md frontmatter and 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\nFile v1.2.0:schema/review-input.example.json\n\n{\n \"period\": {\n \"start\": \"2026-07-13\",\n \"end\": \"2026-07-19\"\n },\n \"source\": {\n \"label\": \"agent-collected\",\n \"note\": \"由运行本 skill 的 Agent 从本平台会话/用量中整理，非 skill 直接读库\"\n },\n \"dashboard\": {\n \"wall_clock_hours\": 42.5,\n \"active_hours\": 18.2,\n \"credits\": 12.4,\n \"session_count\": 27,\n \"avg_session_hours\": 0.67,\n \"long_session_count\": 2,\n \"overnight_idle_count\": 1,\n \"top_models\": [\n { \"name\": \"composer\", \"count\": 15 },\n { \"name\": \"gpt-5\", \"count\": 8 }\n ]\n },\n \"projects\": [\n {\n \"name\": \"weekly-review-skill\",\n \"sessions\": 8,\n \"hours\": 6.5,\n \"credits\": 3.2,\n \"notes\": \"完成 ClawHub 发布与跨平台文案\"\n },\n {\n \"name\": \"TubePilot\",\n \"sessions\": 5,\n \"hours\": 4.1,\n \"credits\": 2.0,\n \"notes\": \"\"\n }\n ],\n \"problems\": [\n {\n \"description\": \"约 1/3 指令仍为动词+无交付物\",\n \"category\": \"【用户】\",\n \"root_cause\": \"提问习惯\",\n \"suggestion\": \"反问训练机制已写入记忆，持续执行\"\n }\n ],\n \"actions\": {\n \"done\": [\n {\n \"action\": \"长会话处置改为先读内容再判断\",\n \"trigger\": \"误判后纠正\",\n \"change\": \"只看时长会错杀合理分工\"\n }\n ],\n \"observing\": [\n {\n \"action\": \"反问训练机制上线\",\n \"criteria\": \"下周模糊指令占比是否下降\"\n }\n ],\n \"pending\": [\n {\n \"suggestion\": \"跨夜 idle 会话下周继续对齐\",\n \"priority\": \"中\",\n \"deadline\": \"2026-07-26\",\n \"status\": \"待开始\"\n }\n ]\n },\n \"open_sessions\": [\n {\n \"title\": \"探索评估 skill 质量的工具\",\n \"created\": \"07/14\",\n \"last_active\": \"07/16\",\n \"cwd_tail\": \"eval-tools\",\n \"kind\": \"cross_week\",\n \"decision\": \"待拍板\",\n \"follow_up\": \"关闭/迁移/继续\"\n }\n ],\n \"automations\": {\n \"runs_total\": 4,\n \"runs_success\": 3,\n \"items\": [\n {\n \"title\": \"周日复盘提醒\",\n \"status\": \"completed\",\n \"result\": \"成功\",\n \"source\": \"—\"\n }\n ]\n }\n}\n\nArchive v1.1.2: 9 files, 19403 bytes\n\nFiles: analyzer.py (10649b), cli.py (3350b), mcp_server.py (6540b), README.md (2770b), report.py (8614b), skill-card.md (2581b), SKILL.md (8061b), utils.py (2757b), _meta.json (132b)\n\nFile v1.1.2:SKILL.md\n\n---\nname: weekly-review\ndescription: >\n Agent-agnostic weekly retrospective from any compatible local AI session\n SQLite DB (not tied to WorkBuddy or a single product): dashboard, per-project\n breakdown, root-cause triage, action ledger, open sessions. Installable on\n OpenClaw / Cursor / Claude Code / any MCP-capable agent. Use for weekly-review\n / 周度复盘 / 本周复盘 / 一周总结.\nversion: 1.1.2\ncategory: 办公效率\nread_when:\n - 用户说\"做周度复盘 / 本周复盘 / 跑一下 weekly review\"\n - 用户想清理或对齐跨周、跨夜还开着的会话\n - 用户希望把复盘流程标准化、沉淀成可复用模板\n - 每周固定时间（如周日）触发复盘\nmetadata:\n openclaw:\n requires:\n bins:\n - python\n emoji: \"📊\"\n homepage: https://github.com/testman2025/weekly-review-skill\n os:\n - macos\n - linux\n - windows\n---\n\n# weekly-review（周度复盘 skill）\n\n一个**与具体 Agent 产品无关**的周度复盘 skill。可安装到 OpenClaw、Cursor、Claude Code，或任何支持 Agent Skill / MCP 的运行时。它把「复盘该看什么、怎么归因、怎么落台账」固化成可复用底座；每周具体项目、人工改进项由使用者或运行时输入。\n\n**不是 WorkBuddy 专用。** WorkBuddy 会话库只是自动发现路径之一；你可以用 `--db-path` / 环境变量 `WEEKLY_REVIEW_DB` 指向任意兼容 schema 的本地 SQLite。\n\n## 适用场景\n\n- **每周定期复盘**，想要客观数据（真实活跃时长、会话数、Credit 消耗、跨周/跨夜会话）支撑，而不是拍脑袋。\n- **会话 / 工作区越开越多**，需要清理或对齐跨周、跨夜仍处于 idle 的长会话。\n- **复盘流程标准化**：团队或个人想把\"复盘该看什么、怎么归因\"沉淀为可复用模板，减少每次从头设计。\n\n## 不适用场景\n\n- **实时会话监控 / 告警**：本 skill 是离线周度分析，不常驻、不主动推送。\n- **非会话类数据复盘**：如代码审查质量、财务报表、用户行为分析——它只吃「会话 / 工作记录」数据。\n- **替你写公众号文章或周报正文**：本 skill 只产出结构化复盘底座，成品内容由你写。\n- **深度主观根因挖掘**：只提供「思路 / 记忆 / 流程」三类归因框架，最终结论要你拍板。\n- **任意专有二进制格式的会话导出**：需要先落到兼容的 SQLite（见下方「数据源约定」），或由 agent 自行整理后走 `--notes`。\n\n## 设计原则\n\n**固化底座，不固化成品。**\n\n- 固化：一页看板、项目分布、根因三类归因、动作台账、待对齐开放会话、自动化概览。\n- 不固化：每周具体项目名、人工标注问题与改进项（由 `--notes` 传入或对话中补充）。\n\n## 安装\n\n本 skill **自包含**：运行代码（各 `.py` 模块）已随本目录一起提供，仅依赖 Python 标准库，**无需 `pip install`、无需额外下载**。运行环境需 Python ≥ 3.10。只读本地 SQLite 会话库，无网络请求、无密钥环境变量。\n\n可安装到**任意**支持 Skill / MCP 的 Agent（不限某一厂商）。\n\n### 从 ClawHub 安装（推荐）\n\n```bash\nopenclaw skills install weekly-review\n# 或\nclawhub install weekly-review\n```\n\n安装后将所用 agent 指向已安装的 skill 目录即可。\n\n### 手动安装\n\n1. 将本 `weekly-review/` 目录（含 `SKILL.md` 与 `cli.py` / `mcp_server.py` / `analyzer.py` / `report.py` / `utils.py`）整体放入所用 agent 的 skills 目录，目录名保持 `weekly-review`。\n2. 确保运行环境有 Python ≥ 3.10。\n3. （可选）设置 `WEEKLY_REVIEW_DB` 指向你的会话库，或运行时传 `--db-path`。\n\n## 使用方式\n\n设 skill 目录为 `{SKILL_DIR}`（即放置本 `SKILL.md` 的 `weekly-review/` 目录）。\n\n### 方式一：CLI 直接生成报告\n\n在 skill 目录下执行：\n\n```bash\ncd {SKILL_DIR}\npython -m cli --start {YYYY-MM-DD} --end {YYYY-MM-DD} -o 周度复盘.md\n```\n\n若需在其他目录调用，用 `PYTHONPATH` 指定 skill 目录：\n\n```bash\nPYTHONPATH={SKILL_DIR} python -m cli --start {YYYY-MM-DD} --end {YYYY-MM-DD} -o 周度复盘.md\n```\n\n### 方式二：MCP server\n\n在你的 agent 的 MCP 配置文件中添加（`PYTHONPATH` 指向 skill 目录）：\n\n```json\n{\n \"mcpServers\": {\n \"weekly-review\": {\n \"command\": \"python\",\n \"args\": [\"-m\", \"mcp_server\"],\n \"env\": { \"PYTHONPATH\": \"{SKILL_DIR}\" }\n }\n }\n}\n```\n\n然后在连接器 / MCP 管理中信任该 server 即可。\n\n### 方式三：对话中调用\n\n当用户说\"做本周复盘\"时，执行：\n\n```bash\nPYTHONPATH={SKILL_DIR} python -m cli --start {本周一} --end {本周日} -o {项目复盘目录}/周度复盘/YYYY-MM-DD周度复盘.md\n```\n\n并向用户展示关键数据，然后用 `AskUserQuestion` 让用户拍板跨周 / 跨夜会话处置。\n\n> 注：若你是从 GitHub 仓库 `testman2025/weekly-review-skill` 安装，进入 `skills/weekly-review/` 后执行 `python -m cli` 即可，无需 `pip install`；本 SKILL.md 以\"随 skill 自带、开箱即用\"为默认路径。\n\n## 输出章节\n\n1. 一页看板\n2. 分项目分析\n3. 问题清单 + 根因三类归因（思路 / 记忆 / 流程）\n4. 本周动作台账（当场已改 / 待观察 / 待落实）\n5. 待对齐开放会话\n6. 自动化运行概览\n\n## 输入\n\n| 输入 | 说明 | 必填 |\n|------|------|------|\n| `--db-path` | 本地 AI 会话库（SQLite）路径 | 否 |\n| `WEEKLY_REVIEW_DB` | 环境变量形式的会话库路径（与 `--db-path` 二选一即可） | 否 |\n| `--start` / `--end` | 统计周期（`YYYY-MM-DD`）；默认上周一～上周日（UTC） | 否 |\n| `--notes` | 人工标注的问题/动作台账 JSON（根因与改进建议由你填写） | 否 |\n| `--output` / `-o` | Markdown 报告输出路径 | 否 |\n| `--json` | 输出原始 JSON 而非 Markdown | 否 |\n\n未指定路径时，会自动探测多种常见位置（含但不限于 WorkBuddy、OpenClaw、Cursor、通用 `sessions.db`）。\n\n## 数据源约定（Agent 无关）\n\n**必需表 `sessions`**（字段名兼容下列列即可）：\n\n- `id`, `cwd`, `title`, `custom_title`, `status`\n- `created_at`, `updated_at`, `last_activity_at`（毫秒或秒时间戳）\n- `model`, `source_mode`, `is_background_automation`\n\n**可选表**（不存在则跳过对应指标，不报错）：\n\n- `session_usage`：`session_id`, `used`, `size`, `credit_json`\n- `automations` / `automation_runs`：自动化定义与运行记录\n\n任何 Agent / 工具只要把会话导出或同步成上述 SQLite，即可复盘；不必使用 WorkBuddy。\n\n## 输出\n\nMarkdown（或 JSON）结构化周度复盘报告，固定包含：一页看板、分项目分析、根因三类归因、动作台账、待对齐开放会话、自动化运行概览。\n\n## 其他约束\n\n- 只读本地 SQLite；不写库、不发起网络请求、不依赖第三方 pip 包。\n- **运行时与数据源均不绑定单一 Agent 产品**；WorkBuddy 仅为可选自动发现路径之一。\n- 报告中的根因分析与改进建议需你通过 `--notes` 或对话补充；工具负责客观数据与格式底座。\n- 在 ClawHub 上发布的 skill 按平台统一许可（MIT-0）分发；本仓库源码许可证见根目录 `README.md`。\n- 长会话处置必须先读会话内容 / 标题推断用途，不能只看时长。\n- 归因禁止一刀切「收敛到单一工作区」，必须区分思路 / 记忆 / 流程三类根因。\n- 跨周会话阈值 48h；跨夜 idle 阈值次日 06:00。\n\n## 操作流程\n\n1. 将本 skill 安装到所用 Agent（ClawHub 或手动复制目录）。\n2. 确认本机有 Python ≥ 3.10，以及可读的兼容会话库（`--db-path` / `WEEKLY_REVIEW_DB` / 自动发现）。\n3. 用 CLI / MCP / 对话触发，指定周期（或使用默认上周）。\n4. 审阅报告中的看板与开放会话，用人工 notes 补齐问题与动作台账。\n5. 对跨周 / 跨夜会话做处置决策（关闭 / 续作 / 归档），并记下待落实项。\n\nFile v1.1.2:README.md\n\n# Weekly Review Skill\n\nAn **agent-agnostic** skill that turns a **local AI session history** (compatible SQLite) into a structured weekly retrospective: one-page dashboard, per-project breakdown, three-class root-cause attribution (approach / memory / process), an action ledger, and open cross-week / overnight sessions to reconcile.\n\nIt works with **any agent runtime** that can host Skills or MCP (OpenClaw, Cursor, Claude Code, etc.). It is **not WorkBuddy-only** — WorkBuddy’s DB path is merely one auto-discovery candidate.\n\nIt **vendors the retrospective \"base\"** and leaves weekly specifics (project names, manual notes) to the caller.\n\n## What it does\n\n- Reads a local AI session DB (SQLite) **read-only**.\n- Produces: one-page dashboard, per-project analysis, problem list with root-cause triage, action ledger, open sessions to align, automation run overview (if tables exist).\n- Exposes three forms: **CLI**, **MCP server** (stdio JSON-RPC), and **Agent SKILL**.\n\n## Safety notes (for moderators / reviewers)\n\n- **No third-party dependencies.** Python standard library only.\n- **No network calls.** Opens a local SQLite file you point at, then prints Markdown/JSON.\n- **Read-only.** Never mutates the source database.\n- **No secrets / env credentials** required (optional `WEEKLY_REVIEW_DB` is only a path).\n- **Agent-agnostic.** Not locked to WorkBuddy or any single vendor.\n\n## Install\n\n```bash\nopenclaw skills install weekly-review\n# or\nclawhub install weekly-review\n```\n\nRequires Python ≥ 3.10. Point any agent at the installed skill directory.\n\n## Data source (not product-specific)\n\n| Priority | How |\n|----------|-----|\n| 1 | `--db-path /path/to/sessions.db` |\n| 2 | Env `WEEKLY_REVIEW_DB` (or `AI_SESSION_DB`) |\n| 3 | Auto-discover common paths (WorkBuddy / OpenClaw / Cursor / `sessions.db`, etc.) |\n\n**Required table:** `sessions` (see `SKILL.md` 「数据源约定」). \n**Optional:** `session_usage`, `automations`, `automation_runs` — skipped if missing.\n\n## Usage\n\n### CLI\n\n```bash\ncd \npython -m cli --db-path /path/to/your-sessions.db --start 2026-07-13 --end 2026-07-19 -o weekly-report.md\n```\n\nOptions: `--db-path`, `--start`, `--end`, `--output / -o`, `--notes`, `--json`.\n\n### MCP server\n\n```json\n{\n \"mcpServers\": {\n \"weekly-review\": {\n \"command\": \"python\",\n \"args\": [\"-m\", \"mcp_server\"],\n \"env\": {\n \"PYTHONPATH\": \"\",\n \"WEEKLY_REVIEW_DB\": \"/path/to/your-sessions.db\"\n }\n }\n }\n}\n```\n\nTool: `run_weekly_review`.\n\n### As an Agent Skill\n\nCopy this directory into your agent’s skills folder (any agent that reads `SKILL.md`).\n\n## License\n\n- Source in this repository: MIT (see repo root).\n- Skills published on ClawHub are distributed under the registry's MIT-0 terms.\n\nFile v1.1.2:_meta.json\n\n{\n \"ownerId\": \"kn71svn18t9cb73emms0vf78s984tb7p\",\n \"slug\": \"weekly-review\",\n \"version\": \"1.1.2\",\n \"publishedAt\": 1784599280707\n}\n\nFile v1.1.2:skill-card.md\n\n## Description:
\nAgent-agnostic weekly retrospective from any compatible local AI session SQLite database: dashboard, per-project breakdown, root-cause triage, action ledger, and open-session reconciliation.
\n\nThis skill is ready for commercial/non-commercial use.
\n\n## Publisher:
\n[testman2025](https://clawhub.ai/user/testman2025)
\n\n### License/Terms of Use:
\nMIT-0
\n\n\n## Use Case:
\nDevelopers, engineers, and teams use this skill to generate offline weekly retrospectives from compatible local AI session SQLite databases, then review project distribution, long-running sessions, root-cause categories, action items, and automation summaries.
\n\n### Deployment Geography for Use:
\nGlobal
\n\n## Known Risks and Mitigations:
\nRisk: The skill reads local AI session metadata, which can include project paths, session titles, usage data, and automation summaries.
\nMitigation: Use the skill only with databases intended for retrospective review and inspect generated reports before sharing them.
\nRisk: Automatic database discovery may select a common local session database when no path is provided.
\nMitigation: Set --db-path or WEEKLY_REVIEW_DB to the intended SQLite file before running the CLI or MCP server.
\nRisk: Root-cause conclusions and action ledgers depend on user-supplied notes and human review.
\nMitigation: Treat the output as a retrospective base and have a human confirm conclusions and open-session dispositions.
\n\n\n## Reference(s):
\n- [ClawHub skill page](https://clawhub.ai/testman2025/skills/weekly-review)
\n- [Publisher profile](https://clawhub.ai/user/testman2025)
\n- [Project homepage](https://github.com/testman2025/weekly-review-skill)
\n- [README.md](artifact/README.md)
\n- [SKILL.md](artifact/SKILL.md)
\n\n\n## Skill Output:
\n**Output Type(s):** [Analysis, Markdown, JSON, Guidance]
\n**Output Format:** [Markdown report or JSON metrics returned by CLI or MCP tool output]
\n**Output Parameters:** [1D]
\n**Other Properties Related to Output:** [Reads a local SQLite database read-only; output can be written to a Markdown file with --output or returned through the MCP tool.]
\n\n## Skill Version(s):
\n1.1.2 (source: release evidence and SKILL.md frontmatter)
\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 v1.1.1: 9 files, 17706 bytes\n\nFiles: analyzer.py (10075b), cli.py (3196b), mcp_server.py (6158b), README.md (3630b), report.py (8415b), skill-card.md (2176b), SKILL.md (6340b), utils.py (1543b), _meta.json (132b)\n\nFile v1.1.1:SKILL.md\n\n---\nname: weekly-review\ndescription: >\n Weekly retrospective from a local AI session DB: one-page dashboard,\n per-project breakdown, three-class root-cause triage (approach/memory/process),\n action ledger, and open cross-week/overnight sessions. Use when the user asks\n for weekly-review / 周度复盘 / 本周复盘 / 一周总结. Vendors the retrospective\n base; weekly specifics stay with the caller.\nversion: 1.1.1\ncategory: 办公效率\nread_when:\n - 用户说\"做周度复盘 / 本周复盘 / 跑一下 weekly review\"\n - 用户想清理或对齐跨周、跨夜还开着的会话\n - 用户希望把复盘流程标准化、沉淀成可复用模板\n - 每周固定时间（如周日）触发复盘\nmetadata:\n openclaw:\n requires:\n bins:\n - python\n emoji: \"📊\"\n homepage: https://github.com/testman2025/weekly-review-skill\n os:\n - macos\n - linux\n - windows\n---\n\n# weekly-review（周度复盘 skill）\n\n一个通用的周度复盘 skill，与具体 AI 平台无关。它把\"复盘该看什么、怎么归因、怎么落台账\"固化成可复用的底座；每周具体项目、人工改进项由使用者或运行时输入。\n\n## 适用场景\n\n- **每周定期复盘**，想要客观数据（真实活跃时长、会话数、Credit 消耗、跨周/跨夜会话）支撑，而不是拍脑袋。\n- **会话 / 工作区越开越多**，需要清理或对齐跨周、跨夜仍处于 idle 的长会话。\n- **复盘流程标准化**：团队或个人想把\"复盘该看什么、怎么归因\"沉淀为可复用模板，减少每次从头设计。\n\n## 不适用场景\n\n- **实时会话监控 / 告警**：本 skill 是离线周度分析，不常驻、不主动推送。\n- **非会话类数据复盘**：如代码审查质量、财务报表、用户行为分析——它只吃\"会话 / 工作记录\"数据。\n- **替你写公众号文章或周报正文**：本 skill 只产出结构化复盘底座，成品内容由你写。\n- **深度主观根因挖掘**：只提供\"思路 / 记忆 / 流程\"三类归因框架，最终结论要你拍板。\n\n## 设计原则\n\n**固化底座，不固化成品。**\n\n- 固化：一页看板、项目分布、根因三类归因、动作台账、待对齐开放会话、自动化概览。\n- 不固化：每周具体项目名、人工标注问题与改进项（由 `--notes` 传入或对话中补充）。\n\n## 安装\n\n本 skill **自包含**：运行代码（各 `.py` 模块）已随本目录一起提供，仅依赖 Python 标准库，**无需 `pip install`、无需额外下载**。运行环境需 Python ≥ 3.10。只读本地 SQLite 会话库，无网络请求、无密钥环境变量。\n\n### 从 ClawHub 安装（推荐）\n\n```bash\nopenclaw skills install weekly-review\n# 或\nclawhub install weekly-review\n```\n\n安装后将 agent 指向已安装的 skill 目录即可。\n\n### 手动安装\n\n1. 将本 `weekly-review/` 目录（含 `SKILL.md` 与 `cli.py` / `mcp_server.py` / `analyzer.py` / `report.py` / `utils.py`）整体放入所用 agent 的 skills 目录，目录名保持 `weekly-review`。\n2. 确保运行环境有 Python ≥ 3.10。\n\n## 使用方式\n\n设 skill 目录为 `{SKILL_DIR}`（即放置本 `SKILL.md` 的 `weekly-review/` 目录）。\n\n### 方式一：CLI 直接生成报告\n\n在 skill 目录下执行：\n\n```bash\ncd {SKILL_DIR}\npython -m cli --start {YYYY-MM-DD} --end {YYYY-MM-DD} -o 周度复盘.md\n```\n\n若需在其他目录调用，用 `PYTHONPATH` 指定 skill 目录：\n\n```bash\nPYTHONPATH={SKILL_DIR} python -m cli --start {YYYY-MM-DD} --end {YYYY-MM-DD} -o 周度复盘.md\n```\n\n### 方式二：MCP server\n\n在你的 agent 的 MCP 配置文件中添加（`PYTHONPATH` 指向 skill 目录）：\n\n```json\n{\n \"mcpServers\": {\n \"weekly-review\": {\n \"command\": \"python\",\n \"args\": [\"-m\", \"mcp_server\"],\n \"env\": { \"PYTHONPATH\": \"{SKILL_DIR}\" }\n }\n }\n}\n```\n\n然后在连接器 / MCP 管理中信任该 server 即可。\n\n### 方式三：对话中调用\n\n当用户说\"做本周复盘\"时，执行：\n\n```bash\nPYTHONPATH={SKILL_DIR} python -m cli --start {本周一} --end {本周日} -o {项目复盘目录}/周度复盘/YYYY-MM-DD周度复盘.md\n```\n\n并向用户展示关键数据，然后用 `AskUserQuestion` 让用户拍板跨周 / 跨夜会话处置。\n\n> 注：若你是从 GitHub 仓库 `testman2025/weekly-review-skill` 安装，进入 `skills/weekly-review/` 后执行 `python -m cli` 即可，无需 `pip install`；本 SKILL.md 以\"随 skill 自带、开箱即用\"为默认路径。\n\n## 输出章节\n\n1. 一页看板\n2. 分项目分析\n3. 问题清单 + 根因三类归因（思路 / 记忆 / 流程）\n4. 本周动作台账（当场已改 / 待观察 / 待落实）\n5. 待对齐开放会话\n6. 自动化运行概览\n\n## 输入\n\n| 输入 | 说明 | 必填 |\n|------|------|------|\n| `--db-path` | 本地 AI 会话库（SQLite）路径；默认自动发现（如 `~/.workbuddy/workbuddy.db`） | 否 |\n| `--start` / `--end` | 统计周期（`YYYY-MM-DD`）；默认上周一～上周日（UTC） | 否 |\n| `--notes` | 人工标注的问题/动作台账 JSON | 否 |\n| `--output` / `-o` | Markdown 报告输出路径 | 否 |\n| `--json` | 输出原始 JSON 而非 Markdown | 否 |\n\n## 输出\n\nMarkdown（或 JSON）结构化周度复盘报告，固定包含：一页看板、分项目分析、根因三类归因、动作台账、待对齐开放会话、自动化运行概览。\n\n## 其他约束\n\n- 只读本地 SQLite；不写库、不发起网络请求、不依赖第三方 pip 包。\n- 在 ClawHub 上发布的 skill 按平台统一许可（MIT-0）分发；本仓库源码许可证见根目录 `README.md`。\n- 长会话处置必须先读会话内容 / 标题推断用途，不能只看时长。\n- 归因禁止一刀切「收敛到单一工作区」，必须区分思路 / 记忆 / 流程三类根因。\n- 跨周会话阈值 48h；跨夜 idle 阈值次日 06:00。\n\n## 操作流程\n\n1. 安装 skill（ClawHub 或手动复制本目录）。\n2. 确认本机有 Python ≥ 3.10，以及可读的本地会话库（或准备好 `--notes`）。\n3. 用 CLI / MCP / 对话触发，指定周期（或使用默认上周）。\n4. 审阅报告中的看板与开放会话，用人工 notes 补齐问题与动作台账。\n5. 对跨周 / 跨夜会话做处置决策（关闭 / 续作 / 归档），并记下待落实项。\n\nFile v1.1.1:README.md\n\n# Weekly Review Skill\n\nA self-contained, platform-agnostic skill that turns your **local AI session history** (a SQLite database) into a structured weekly retrospective: a one-page dashboard, per-project breakdown, three-class root-cause attribution (approach / memory / process), an action ledger, and a list of open cross-week / overnight sessions to reconcile.\n\nIt **vendors the retrospective \"base\"** (what to look at, how to attribute causes, how to log actions) and leaves weekly specifics (project names, manual notes) to the caller.\n\n## What it does\n\n- Reads a local AI session DB (SQLite) **read-only**.\n- Produces: one-page dashboard (wall-clock, real active time, credit usage, session count, long/overnight sessions), per-project analysis, problem list with root-cause triage, action ledger (done / observing / pending), open sessions to align, automation run overview.\n- Exposes three forms: **CLI**, **MCP server** (stdio JSON-RPC), and **Agent SKILL**.\n\n## Safety notes (for moderators / reviewers)\n\n- **No third-party dependencies.** Uses only the Python standard library (`sqlite3`, `json`, `argparse`, `datetime`, `pathlib`, `typing`). There is no `pip install` step and nothing is downloaded at runtime.\n- **No network calls.** The skill only opens a local SQLite file you point it at (default `~/.workbuddy/workbuddy.db`), reads it, and prints a Markdown/JSON report to stdout or a local file.\n- **Read-only.** It never writes to or mutates the source database.\n- **No secrets / env credentials.** No API keys or tokens required.\n- **No hidden execution.** The only \"code\" that runs is the bundled Python scripts, executed locally on your own machine, on data you explicitly provide.\n\n## Install\n\n### From ClawHub (recommended)\n\n```bash\nopenclaw skills install weekly-review\n# or\nclawhub install weekly-review\n```\n\nThen point your agent at the installed skill directory. Requires Python ≥ 3.10 on the host.\n\n### Manual\n\nCopy this entire `weekly-review/` folder (including `SKILL.md` and all `.py` files) into your agent's skills directory.\n\n## Usage\n\n### CLI\n\n```bash\ncd \npython -m cli --start 2026-07-13 --end 2026-07-19 -o weekly-report.md\n```\n\nOptions: `--db-path`, `--start`, `--end`, `--output / -o`, `--notes` (JSON of manual problems/actions), `--json`.\n\n### MCP server\n\n```json\n{\n \"mcpServers\": {\n \"weekly-review\": {\n \"command\": \"python\",\n \"args\": [\"-m\", \"mcp_server\"],\n \"env\": { \"PYTHONPATH\": \"\" }\n }\n }\n}\n```\n\nTool: `run_weekly_review` (params: `db_path`, `start_date`, `end_date`, `output_format`, `notes`).\n\n### As an Agent Skill\n\nCopy this directory's `SKILL.md` into your agent's skills folder (e.g. `~/.agents/skills/weekly-review/SKILL.md`).\n\n## Output sections\n\n1. One-page dashboard\n2. Per-project analysis (auto-aggregated by working-directory leaf)\n3. Problem list + three-class root-cause attribution (approach / memory / process)\n4. Action ledger (done / observing / pending)\n5. Open sessions to reconcile (>48h cross-week, overnight idle)\n6. Automation run overview\n\n## Notes\n\n- Default period: last Monday 00:00 ~ last Sunday 23:59 (UTC), overridable via `--start` / `--end`.\n- Cross-week threshold: 48h. Overnight idle threshold: last active after 06:00 next day.\n- Long sessions must be reconciled by reading their content/title, not by duration alone.\n- Root-cause attribution must distinguish approach / memory / process — never a one-size \"collapse everything into one workspace\" verdict.\n\n## License\n\n- Source in this repository: MIT (see repo root).\n- Skills published on ClawHub are distributed under the registry's MIT-0 terms.\n\nFile v1.1.1:_meta.json\n\n{\n \"ownerId\": \"kn71svn18t9cb73emms0vf78s984tb7p\",\n \"slug\": \"weekly-review\",\n \"version\": \"1.1.1\",\n \"publishedAt\": 1784543444875\n}\n\nFile v1.1.1:skill-card.md\n\n## Description:
\nWeekly Review turns a local AI session SQLite database into a structured weekly retrospective with a one-page dashboard, per-project breakdown, root-cause triage, action ledger, and open cross-week or overnight sessions to reconcile.
\n\nThis skill is ready for commercial/non-commercial use.
\n\n## Publisher:
\n[testman2025](https://clawhub.ai/user/testman2025)
\n\n### License/Terms of Use:
\nMIT-0
\n\n\n## Use Case:
\nDevelopers, operators, and individual AI-tool users use this skill to summarize weekly local AI session activity, identify long-running or overnight sessions, and prepare a concise retrospective with follow-up actions.
\n\n### Deployment Geography for Use:
\nGlobal
\n\n## Known Risks and Mitigations:
\nRisk: The generated report may expose sensitive local session details such as session titles, project folder names, model usage, credits, and automation summaries.
\nMitigation: Run it only against databases you intend to review, pass an explicit --db-path, choose a clear local --output path, and inspect the report before sharing it.
\n\n\n## Reference(s):
\n- [Weekly Review on ClawHub](https://clawhub.ai/testman2025/skills/weekly-review)
\n- [Project homepage](https://github.com/testman2025/weekly-review-skill)
\n\n\n## Skill Output:
\n**Output Type(s):** [text, markdown, json, shell commands, configuration, guidance]
\n**Output Format:** [Markdown or JSON weekly retrospective report, with optional CLI commands and MCP configuration guidance]
\n**Output Parameters:** [1D]
\n**Other Properties Related to Output:** [Reports are generated from a local read-only SQLite session database and may include session titles, project folder names, model usage, credit usage, and automation summaries.]
\n\n## Skill Version(s):
\n1.1.1 (source: server release evidence and skill frontmatter)
\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 v1.1.0: 9 files, 16734 bytes\n\nFiles: analyzer.py (10075b), cli.py (3196b), mcp_server.py (6158b), README.md (3190b), report.py (8415b), skill-card.md (2343b), SKILL.md (4666b), utils.py (1543b), _meta.json (132b)\n\nFile v1.1.0:SKILL.md\n\n---\nname: weekly-review\ndescription: \"通用周度复盘 skill：读取本地 AI 会话库，基于会话时长、活跃度、自动化运行等客观数据，自动产出一页看板、分项目分析、根因三类归因（思路/记忆/流程）、动作台账、待对齐开放会话。固化复盘底座，不固化每周成品。\"\nversion: 1.1.0\ncategory: 办公效率\nread_when:\n - 用户说\"做周度复盘 / 本周复盘 / 跑一下 weekly review\"\n - 用户想清理或对齐跨周、跨夜还开着的会话\n - 用户希望把复盘流程标准化、沉淀成可复用模板\n - 每周固定时间（如周日）触发复盘\n---\n\n# weekly-review（周度复盘 skill）\n\n一个通用的周度复盘 skill，与具体 AI 平台无关。它把\"复盘该看什么、怎么归因、怎么落台账\"固化成可复用的底座；每周具体项目、人工改进项由使用者或运行时输入。\n\n## 适用场景\n\n- **每周定期复盘**，想要客观数据（真实活跃时长、会话数、Credit 消耗、跨周/跨夜会话）支撑，而不是拍脑袋。\n- **会话 / 工作区越开越多**，需要清理或对齐跨周、跨夜仍处于 idle 的长会话。\n- **复盘流程标准化**：团队或个人想把\"复盘该看什么、怎么归因\"沉淀为可复用模板，减少每次从头设计。\n\n## 不适用场景\n\n- **实时会话监控 / 告警**：本 skill 是离线周度分析，不常驻、不主动推送。\n- **非会话类数据复盘**：如代码审查质量、财务报表、用户行为分析——它只吃\"会话 / 工作记录\"数据。\n- **替你写公众号文章或周报正文**：本 skill 只产出结构化复盘底座，成品内容由你写。\n- **深度主观根因挖掘**：只提供\"思路 / 记忆 / 流程\"三类归因框架，最终结论要你拍板。\n\n## 设计原则\n\n**固化底座，不固化成品。**\n\n- 固化：一页看板、项目分布、根因三类归因、动作台账、待对齐开放会话、自动化概览。\n- 不固化：每周具体项目名、人工标注问题与改进项（由 `--notes` 传入或对话中补充）。\n\n## 安装\n\n本 skill **自包含**：运行代码（各 `.py` 模块）已随本目录一起提供，仅依赖 Python 标准库，**无需 `pip install`、无需额外下载**。\n\n1. 将本 `weekly-review/` 目录（含 `SKILL.md` 与 `cli.py` / `mcp_server.py` / `analyzer.py` / `report.py` / `utils.py`）整体放入所用 agent 的 skills 目录，目录名保持 `weekly-review`。\n2. 确保运行环境有 Python ≥ 3.10。\n\n## 使用方式\n\n设 skill 目录为 `{SKILL_DIR}`（即放置本 `SKILL.md` 的 `weekly-review/` 目录）。\n\n### 方式一：CLI 直接生成报告\n\n在 skill 目录下执行：\n\n```bash\ncd {SKILL_DIR}\npython -m cli --start {YYYY-MM-DD} --end {YYYY-MM-DD} -o 周度复盘.md\n```\n\n若需在其他目录调用，用 `PYTHONPATH` 指定 skill 目录：\n\n```bash\nPYTHONPATH={SKILL_DIR} python -m cli --start {YYYY-MM-DD} --end {YYYY-MM-DD} -o 周度复盘.md\n```\n\n### 方式二：MCP server\n\n在你的 agent 的 MCP 配置文件中添加（`PYTHONPATH` 指向 skill 目录）：\n\n```json\n{\n \"mcpServers\": {\n \"weekly-review\": {\n \"command\": \"python\",\n \"args\": [\"-m\", \"mcp_server\"],\n \"env\": { \"PYTHONPATH\": \"{SKILL_DIR}\" }\n }\n }\n}\n```\n\n然后在连接器 / MCP 管理中信任该 server 即可。\n\n### 方式三：对话中调用\n\n当用户说\"做本周复盘\"时，执行：\n\n```bash\nPYTHONPATH={SKILL_DIR} python -m cli --start {本周一} --end {本周日} -o {项目复盘目录}/周度复盘/YYYY-MM-DD周度复盘.md\n```\n\n并向用户展示关键数据，然后用 `AskUserQuestion` 让用户拍板跨周 / 跨夜会话处置。\n\n> 注：若你是从 GitHub 仓库 `testman2025/weekly-review-skill` 安装，进入 `skills/weekly-review/` 后执行 `python -m cli` 即可，无需 `pip install`；本 SKILL.md 以\"随 skill 自带、开箱即用\"为默认路径。\n\n## 输出章节\n\n1. 一页看板\n2. 分项目分析\n3. 问题清单 + 根因三类归因（思路 / 记忆 / 流程）\n4. 本周动作台账（当场已改 / 待观察 / 待落实）\n5. 待对齐开放会话\n6. 自动化运行概览\n\n## 注意事项\n\n- 统计周期默认上周一 00:00 ~ 上周日 23:59（UTC），可用 `--start` / `--end` 覆盖。\n- 默认读取本地 AI 会话库；可用 `--db-path` 指定其他会话库，用 `--notes` 传入自定义复盘笔记 JSON。\n- 跨周会话阈值 48h，跨夜 idle 阈值次日 06:00。\n- 长会话处置必须先读会话内容 / 标题推断用途，不能只看时长。\n- 归因禁止一刀切\"收敛到单一工作区\"，必须区分思路 / 记忆 / 流程三类根因。\n\nFile v1.1.0:README.md\n\n# Weekly Review Skill\n\nA self-contained, platform-agnostic skill that turns your **local AI session history** (a SQLite database) into a structured weekly retrospective: a one-page dashboard, per-project breakdown, three-class root-cause attribution (approach / memory / process), an action ledger, and a list of open cross-week / overnight sessions to reconcile.\n\nIt **vendors the retrospective \"base\"** (what to look at, how to attribute causes, how to log actions) and leaves weekly specifics (project names, manual notes) to the caller.\n\n## What it does\n\n- Reads a local AI session DB (SQLite) **read-only**.\n- Produces: one-page dashboard (wall-clock, real active time, credit usage, session count, long/overnight sessions), per-project analysis, problem list with root-cause triage, action ledger (done / observing / pending), open sessions to align, automation run overview.\n- Exposes three forms: **CLI**, **MCP server** (stdio JSON-RPC), and **Agent SKILL**.\n\n## Safety notes (for moderators / reviewers)\n\n- **No third-party dependencies.** Uses only the Python standard library (`sqlite3`, `json`, `argparse`, `datetime`, `pathlib`, `typing`). There is no `pip install` step and nothing is downloaded at runtime.\n- **No network calls.** The skill only opens a local SQLite file you point it at (default `~/.workbuddy/workbuddy.db`), reads it, and prints a Markdown/JSON report to stdout or a local file.\n- **Read-only.** It never writes to or mutates the source database.\n- **No hidden execution.** The only \"code\" that runs is the bundled Python scripts, executed locally on your own machine, on data you explicitly provide.\n\n## Install\n\n```bash\nclawhub install weekly-review\n```\n\nThen point your agent at the installed skill directory.\n\n## Usage\n\n### CLI\n\n```bash\ncd \npython -m cli --start 2026-07-13 --end 2026-07-19 -o weekly-report.md\n```\n\nOptions: `--db-path`, `--start`, `--end`, `--output / -o`, `--notes` (JSON of manual problems/actions), `--json`.\n\n### MCP server\n\n```json\n{\n \"mcpServers\": {\n \"weekly-review\": {\n \"command\": \"python\",\n \"args\": [\"-m\", \"mcp_server\"],\n \"env\": { \"PYTHONPATH\": \"\" }\n }\n }\n}\n```\n\nTool: `run_weekly_review` (params: `db_path`, `start_date`, `end_date`, `output_format`, `notes`).\n\n### As an Agent Skill\n\nCopy this directory's `SKILL.md` into your agent's skills folder (e.g. `~/.agents/skills/weekly-review/SKILL.md`).\n\n## Output sections\n\n1. One-page dashboard\n2. Per-project analysis (auto-aggregated by working-directory leaf)\n3. Problem list + three-class root-cause attribution (approach / memory / process)\n4. Action ledger (done / observing / pending)\n5. Open sessions to reconcile (>48h cross-week, overnight idle)\n6. Automation run overview\n\n## Notes\n\n- Default period: last Monday 00:00 ~ last Sunday 23:59 (UTC), overridable via `--start` / `--end`.\n- Cross-week threshold: 48h. Overnight idle threshold: last active after 06:00 next day.\n- Long sessions must be reconciled by reading their content/title, not by duration alone.\n- Root-cause attribution must distinguish approach / memory / process — never a one-size \"collapse everything into one workspace\" verdict.\n\n## License\n\nMIT\n\nFile v1.1.0:_meta.json\n\n{\n \"ownerId\": \"kn71svn18t9cb73emms0vf78s984tb7p\",\n \"slug\": \"weekly-review\",\n \"version\": \"1.1.0\",\n \"publishedAt\": 1784542894720\n}\n\nFile v1.1.0:skill-card.md\n\n## Description:
\nWeekly Review reads a local AI session SQLite database and produces a structured weekly retrospective with a dashboard, per-project breakdown, root-cause triage, action ledger, and open-session reconciliation list.
\n\nThis skill is ready for commercial/non-commercial use.
\n\n## Publisher:
\n[testman2025](https://clawhub.ai/user/testman2025)
\n\n### License/Terms of Use:
\nMIT-0
\n\n\n## Use Case:
\nDevelopers, operators, and individual users use this skill to turn local AI session history into a weekly retrospective that supports project review, workflow cleanup, and follow-up action tracking.
\n\n### Deployment Geography for Use:
\nGlobal
\n\n## Known Risks and Mitigations:
\nRisk: Generated reports may include project names, local paths, session titles, timing, and credit metrics from the local session database.
\nMitigation: Use an explicit --db-path and --output location, keep reports local by default, and review report contents before sharing.
\nRisk: Scheduled or cleanup-triggered reviews may read the default local session database when the user did not intend to run a review.
\nMitigation: Configure the agent to ask before running scheduled reviews or acting on long-running and overnight sessions.
\nRisk: Root-cause triage is a retrospective aid and may not fully capture subjective workflow causes.
\nMitigation: Treat the generated triage as a draft, add user-provided notes, and require human confirmation before turning findings into actions.
\n\n\n## Reference(s):
\n\n\n## Skill Output:
\n**Output Type(s):** [text, markdown, json, guidance]
\n**Output Format:** [Markdown or JSON report, returned through CLI stdout, a local output file, or the MCP tool response.]
\n**Output Parameters:** [1D]
\n**Other Properties Related to Output:** [Reads a local SQLite session database and optional notes JSON; default output is a weekly Markdown retrospective.]
\n\n## Skill Version(s):
\n1.1.0 (source: server release metadata and SKILL.md frontmatter)
\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 v1.0.1: 9 files, 15681 bytes\n\nFiles: skill-card.md (2243b), SKILL.md (4631b), weekly_review/__init__.py (537b), weekly_review/analyzer.py (10076b), weekly_review/cli.py (3198b), weekly_review/mcp_server.py (6160b), weekly_review/report.py (8417b), weekly_review/utils.py (1543b), _meta.json (132b)\n\nFile v1.0.1:SKILL.md\n\n---\nname: weekly-review\ndescription: \"通用周度复盘 skill：读取本地 AI 会话库，基于会话时长、活跃度、自动化运行等客观数据，自动产出一页看板、分项目分析、根因三类归因（思路/记忆/流程）、动作台账、待对齐开放会话。固化复盘底座，不固化每周成品。\"\nversion: 1.0.1\ncategory: 办公效率\nread_when:\n - 用户说\"做周度复盘 / 本周复盘 / 跑一下 weekly review\"\n - 用户想清理或对齐跨周、跨夜还开着的会话\n - 用户希望把复盘流程标准化、沉淀成可复用模板\n - 每周固定时间（如周日）触发复盘\n---\n\n# weekly-review（周度复盘 skill）\n\n一个通用的周度复盘 skill，与具体 AI 平台无关。它把\"复盘该看什么、怎么归因、怎么落台账\"固化成可复用的底座；每周具体项目、人工改进项由使用者或运行时输入。\n\n## 适用场景\n\n- **每周定期复盘**，想要客观数据（真实活跃时长、会话数、Credit 消耗、跨周/跨夜会话）支撑，而不是拍脑袋。\n- **会话 / 工作区越开越多**，需要清理或对齐跨周、跨夜仍处于 idle 的长会话。\n- **复盘流程标准化**：团队或个人想把\"复盘该看什么、怎么归因\"沉淀为可复用模板，减少每次从头设计。\n\n## 不适用场景\n\n- **实时会话监控 / 告警**：本 skill 是离线周度分析，不常驻、不主动推送。\n- **非会话类数据复盘**：如代码审查质量、财务报表、用户行为分析——它只吃\"会话 / 工作记录\"数据。\n- **替你写公众号文章或周报正文**：本 skill 只产出结构化复盘底座，成品内容由你写。\n- **深度主观根因挖掘**：只提供\"思路 / 记忆 / 流程\"三类归因框架，最终结论要你拍板。\n\n## 设计原则\n\n**固化底座，不固化成品。**\n\n- 固化：一页看板、项目分布、根因三类归因、动作台账、待对齐开放会话、自动化概览。\n- 不固化：每周具体项目名、人工标注问题与改进项（由 `--notes` 传入或对话中补充）。\n\n## 安装\n\n本 skill **自包含**：运行代码（`weekly_review/` 包）已随本目录一起提供，仅依赖 Python 标准库，**无需 `pip install`、无需额外下载**。\n\n1. 将本 `weekly-review/` 目录（含 `SKILL.md` 与 `weekly_review/`）整体放入所用 agent 的 skills 目录，目录名保持 `weekly-review`。\n2. 确保运行环境有 Python ≥ 3.10。\n\n## 使用方式\n\n设 skill 目录为 `{SKILL_DIR}`（即放置本 `SKILL.md` 的 `weekly-review/` 目录）。\n\n### 方式一：CLI 直接生成报告\n\n在 skill 目录下执行：\n\n```bash\ncd {SKILL_DIR}\npython -m weekly_review.cli --start {YYYY-MM-DD} --end {YYYY-MM-DD} -o 周度复盘.md\n```\n\n或在任意目录通过 `PYTHONPATH` 指定：\n\n```bash\nPYTHONPATH={SKILL_DIR} python -m weekly_review.cli --start {YYYY-MM-DD} --end {YYYY-MM-DD} -o 周度复盘.md\n```\n\n### 方式二：MCP server\n\n在你的 agent 的 MCP 配置文件中添加（`PYTHONPATH` 指向 skill 目录）：\n\n```json\n{\n \"mcpServers\": {\n \"weekly-review\": {\n \"command\": \"python\",\n \"args\": [\"-m\", \"weekly_review.mcp_server\"],\n \"env\": { \"PYTHONPATH\": \"{SKILL_DIR}\" }\n }\n }\n}\n```\n\n然后在连接器 / MCP 管理中信任该 server 即可。\n\n### 方式三：对话中调用\n\n当用户说\"做本周复盘\"时，执行：\n\n```bash\nPYTHONPATH={SKILL_DIR} python -m weekly_review.cli --start {本周一} --end {本周日} -o {项目复盘目录}/周度复盘/YYYY-MM-DD周度复盘.md\n```\n\n并向用户展示关键数据，然后用 `AskUserQuestion` 让用户拍板跨周 / 跨夜会话处置。\n\n> 注：若你是从 GitHub 仓库 `testman2025/weekly-review-skill` 安装并希望走 `pip install -e .` 全局命令，请参见仓库 README；本 SKILL.md 以\"随 skill 自带、开箱即用\"为默认路径。\n\n## 输出章节\n\n1. 一页看板\n2. 分项目分析\n3. 问题清单 + 根因三类归因（思路 / 记忆 / 流程）\n4. 本周动作台账（当场已改 / 待观察 / 待落实）\n5. 待对齐开放会话\n6. 自动化运行概览\n\n## 注意事项\n\n- 统计周期默认上周一 00:00 ~ 上周日 23:59（UTC），可用 `--start` / `--end` 覆盖。\n- 默认读取本地 AI 会话库；可用 `--db-path` 指定其他会话库，用 `--notes` 传入自定义复盘笔记 JSON。\n- 跨周会话阈值 48h，跨夜 idle 阈值次日 06:00。\n- 长会话处置必须先读会话内容 / 标题推断用途，不能只看时长。\n- 归因禁止一刀切\"收敛到单一工作区\"，必须区分思路 / 记忆 / 流程三类根因。\n\nFile v1.0.1:_meta.json\n\n{\n \"ownerId\": \"kn71svn18t9cb73emms0vf78s984tb7p\",\n \"slug\": \"weekly-review\",\n \"version\": \"1.0.1\",\n \"publishedAt\": 1784541977250\n}\n\nFile v1.0.1:skill-card.md\n\n## Description:
\n周度复盘 weekly-review generates structured weekly review reports from local AI session activity, including activity metrics, project breakdowns, root-cause prompts, action logs, open sessions, and automation summaries.
\n\nThis skill is ready for commercial/non-commercial use.
\n\n## Publisher:
\n[testman2025](https://clawhub.ai/user/testman2025)
\n\n### License/Terms of Use:
\nMIT-0
\n\n\n## Use Case:
\nDevelopers, teams, and individual users use this skill to turn local AI session records into a repeatable weekly review baseline. It is intended for offline reflection, project analysis, action tracking, and cleanup decisions for long-running or idle sessions.
\n\n### Deployment Geography for Use:
\nGlobal
\n\n## Known Risks and Mitigations:
\nRisk: The skill reads local AI session activity and may include project names, paths, titles, model usage, credit usage, and automation history in generated reports or MCP responses.
\nMitigation: Install only in environments where that local access is acceptable, choose explicit date ranges, and treat generated reports as potentially sensitive.
\nRisk: The CLI can write reports to a user-specified output path.
\nMitigation: Review the output path before using the output option and store reports in an appropriate location for the sensitivity of the session data.
\n\n\n## Reference(s):
\n- [ClawHub skill page](https://clawhub.ai/testman2025/skills/weekly-review)
\n\n\n## Skill Output:
\n**Output Type(s):** [markdown, text, JSON, shell commands, configuration, guidance]
\n**Output Format:** [Markdown reports, JSON metrics, CLI commands, and MCP tool responses]
\n**Output Parameters:** [1D]
\n**Other Properties Related to Output:** [Can write a Markdown report to a chosen output path or return JSON metrics through CLI or MCP.]
\n\n## Skill Version(s):
\n1.0.1 (source: release evidence and SKILL.md frontmatter)
\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 v1.0.0: 3 files, 3617 bytes\n\nFiles: skill-card.md (1974b), SKILL.md (3771b), _meta.json (132b)\n\nFile v1.0.0:SKILL.md\n\n---\nname: weekly-review\ndescription: \"通用周度复盘 skill：读取本地 AI 会话库，基于会话时长、活跃度、自动化运行等客观数据，自动产出一页看板、分项目分析、根因三类归因（思路/记忆/流程）、动作台账、待对齐开放会话。固化复盘底座，不固化每周成品。\"\nversion: 1.0.0\ncategory: 办公效率\nread_when:\n - 用户说\"做周度复盘 / 本周复盘 / 跑一下 weekly review\"\n - 用户想清理或对齐跨周、跨夜还开着的会话\n - 用户希望把复盘流程标准化、沉淀成可复用模板\n - 每周固定时间（如周日）触发复盘\n---\n\n# weekly-review（周度复盘 skill）\n\n一个通用的周度复盘 skill，与具体 AI 平台无关。它把\"复盘该看什么、怎么归因、怎么落台账\"固化成可复用的底座；每周具体项目、人工改进项由使用者或运行时输入。\n\n## 适用场景\n\n- **每周定期复盘**，想要客观数据（真实活跃时长、会话数、Credit 消耗、跨周/跨夜会话）支撑，而不是拍脑袋。\n- **会话 / 工作区越开越多**，需要清理或对齐跨周、跨夜仍处于 idle 的长会话。\n- **复盘流程标准化**：团队或个人想把\"复盘该看什么、怎么归因\"沉淀为可复用模板，减少每次从头设计。\n\n## 不适用场景\n\n- **实时会话监控 / 告警**：本 skill 是离线周度分析，不常驻、不主动推送。\n- **非会话类数据复盘**：如代码审查质量、财务报表、用户行为分析——它只吃\"会话 / 工作记录\"数据。\n- **替你写公众号文章或周报正文**：本 skill 只产出结构化复盘底座，成品内容由你写。\n- **深度主观根因挖掘**：只提供\"思路 / 记忆 / 流程\"三类归因框架，最终结论要你拍板。\n\n## 设计原则\n\n**固化底座，不固化成品。**\n\n- 固化：一页看板、项目分布、根因三类归因、动作台账、待对齐开放会话、自动化概览。\n- 不固化：每周具体项目名、人工标注问题与改进项（由 `--notes` 传入或对话中补充）。\n\n## 安装\n\n1. 确保本仓库已安装：\n ```bash\n cd weekly-review-skill\n pip install -e .\n ```\n2. 将本 `SKILL.md` 复制到所用 agent 的 skills 目录（目录名 `weekly-review`，放在 skills 根下）。\n\n## 使用方式\n\n### 方式一：CLI 直接生成报告\n\n```bash\nweekly-review --start {YYYY-MM-DD} --end {YYYY-MM-DD} -o 周度复盘.md\n```\n\n### 方式二：MCP server\n\n在你的 agent 的 MCP 配置文件中添加：\n\n```json\n{\n \"mcpServers\": {\n \"weekly-review\": {\n \"command\": \"weekly-review-mcp\"\n }\n }\n}\n```\n\n然后在连接器 / MCP 管理中信任该 server 即可。\n\n### 方式三：对话中调用\n\n当用户说\"做本周复盘\"时，执行：\n\n```bash\nweekly-review --start {本周一} --end {本周日} -o {项目复盘目录}/周度复盘/YYYY-MM-DD周度复盘.md\n```\n\n并向用户展示关键数据，然后用 `AskUserQuestion` 让用户拍板跨周 / 跨夜会话处置。\n\n## 输出章节\n\n1. 一页看板\n2. 分项目分析\n3. 问题清单 + 根因三类归因（思路 / 记忆 / 流程）\n4. 本周动作台账（当场已改 / 待观察 / 待落实）\n5. 待对齐开放会话\n6. 自动化运行概览\n\n## 注意事项\n\n- 统计周期默认上周一 00:00 ~ 上周日 23:59（UTC），可用 `--start` / `--end` 覆盖。\n- 默认读取本地 AI 会话库；可用 `--db-path` 指定其他会话库，用 `--notes` 传入自定义复盘笔记 JSON。\n- 跨周会话阈值 48h，跨夜 idle 阈值次日 06:00。\n- 长会话处置必须先读会话内容 / 标题推断用途，不能只看时长。\n- 归因禁止一刀切\"收敛到单一工作区\"，必须区分思路 / 记忆 / 流程三类根因。\n\nFile v1.0.0:_meta.json\n\n{\n \"ownerId\": \"kn71svn18t9cb73emms0vf78s984tb7p\",\n \"slug\": \"weekly-review\",\n \"version\": \"1.0.0\",\n \"publishedAt\": 1784541495363\n}\n\nFile v1.0.0:skill-card.md\n\n## Description:
\n通用周度复盘 skill：读取本地 AI 会话库，基于会话时长、活跃度、自动化运行等客观数据，自动产出一页看板、分项目分析、根因三类归因（思路/记忆/流程）、动作台账、待对齐开放会话。固化复盘底座，不固化每周成品。
\n\nThis skill is ready for commercial/non-commercial use.
\n\n## Publisher:
\n[testman2025](https://clawhub.ai/user/testman2025)
\n\n### License/Terms of Use:
\nMIT-0
\n\n\n## Use Case:
\nDevelopers, teams, and individual AI tool users use this skill to generate an offline weekly retrospective from local AI session records, including activity metrics, project analysis, root-cause categories, an action ledger, and open-session follow-up items.
\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- [ClawHub skill page](https://clawhub.ai/testman2025/skills/weekly-review)
\n\n\n## Skill Output:
\n**Output Type(s):** [Markdown, Shell commands, Configuration, Guidance]
\n**Output Format:** [Markdown report with inline shell commands and JSON configuration examples]
\n**Output Parameters:** [1D]
\n**Other Properties Related to Output:** [Reads local AI session data and writes a persistent weekly review file; confirm the date range, database path, output location, and any needed redaction before running.]
\n\n## Skill Version(s):
\n1.0.0 (source: frontmatter and server-resolved 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.
","readmeExcerpt":"Skill: Weekly Review Owner: testman2025 Summary: 面向任意 AI Agent 自动复盘的 Skill（weekly-review / 复盘 / 提示词优化 / 会话清理）。 核心能力覆盖：AI 用量与时间看板、提示词复盘与改写建议、高效/低效归因、 时间分布与 Token/归因趋势图、开放会话对齐清理、定时自动周报。 当用户任务涉及周度复盘、用量查看、提示词改进... Tags: agent-agnostic:1.1.2, agent-fed:1.2.3, automation:1.1.0, latest:1.2.5, office:1.1.0, productivity:1.2.1, retrospective:1.2.3, weekly-review:1.2.3 Version history: v1.2.5 | 2026-07-24T03:31:39.454Z | user ","codeSnippets":[],"executableExamples":[{"language":"bash","snippet":"cd {SKILL_DIR}\npython -m cli --input schema/review-input.example.json -o 周度复盘.md\n# 图表默认写到报告同目录；关闭：加 --no-charts"},{"language":"bash","snippet":"npx skills add testman2025/weekly-review-skill --skill weekly-review\n# or\nopenclaw skills install weekly-review"},{"language":"bash","snippet":"python -m cli --input schema/review-input.example.json -o weekly-report.md"},{"language":"json","snippet":"\"charts_data\": {\n \"time_distribution\": [{ \"label\": \"社媒 Agent 主题\", \"hours\": 15.4 }],\n \"attribution\": [{ \"label\": \"【用户】指令模糊\", \"count\": 3 }],\n \"token_trend\": [{ \"label\": \"本周\", \"millions\": 2.74 }]\n}"},{"language":"bash","snippet":"cd {SKILL_DIR}\npython -m cli --input schema/review-input.example.json -o 周度复盘.md\n# 图表默认写到报告同目录；关闭：加 --no-charts"},{"language":"bash","snippet":"npx skills add testman2025/weekly-review-skill --skill weekly-review\n# or\nopenclaw skills install weekly-review"}],"parameters":null,"dependencies":[],"permissions":[],"extractedFiles":[{"path":"SKILL.md","content":"---\nname: weekly-review\ndescription: >\n 面向任意 AI Agent 自动复盘的 Skill（weekly-review / 复盘 / 提示词优化 / 会话清理）。\n\n 核心能力覆盖：AI 用量与时间看板、提示词复盘与改写建议、高效/低效归因、\n 时间分布与 Token/归因趋势图、开放会话对齐清理、定时自动周报。\n\n 当用户任务涉及周度复盘、用量查看、提示词改进时使用，包括但不限于：\n Token 花了多少、提示词优化改进、一周总结、工作复盘、复盘可视化报告、\n 清理对话、AI 自动复盘。\nversion: 1.2.5\ncategory: 办公效率\nread_when:\n - 提示词改进\n - 本周 AI 用量看板\n - 高低效归因\n - 复盘可视化报告\n - 清理开放对话\n - 定时/自动周报\n - weekly-review / 复盘 / /weekly-review\nmetadata:\n openclaw:\n requires:\n bins:\n - python\n emoji: \"📊\"\n homepage: https://github.com/testman2025/weekly-review-skill\n os:\n - macos\n - linux\n - windows\n---\n\n# AI 用量与提示词复盘助手（weekly-review）\n\n用来做：提示词改进、本周 AI 用量看板、高低效归因、复盘可视化图表、开放会话清理、定时/自动周报。宿主 Agent 采数，本 skill 按六章模板出报告。\n\n## 输出结构（锁定）\n\n1. **一页看板**：`| 指标 | 数值 | 口径说明 |` + **一句话结论** + 可选「辅助图表」PNG 路径 \n2. **分项目分析**：`### 2.x 主题（时长/会话）` + `| 维度 | 内容 |`（做了什么 / DB cwd / 磁盘核对 / 低效 / 提示词）；其他用简表 \n3. **问题清单 + 根因三类归因** + **做得好的（正面范本）** \n4. **本周动作台账 ★**：当场已改 / 观察 / 待落实（或待落实→执行结果） \n5. **待对齐开放会话** \n6. **自动化概览** \n7. **事实更正**（可选，有核对推翻时必写）\n\n**图表约定**：渲染器生成两张定稿样式图——横向「时间投入分布」、双栏「低效归因 + Token 趋势」（SVG）。数据来自 `charts_data`。看板主表仍是 `| 指标 | 数值 | 口径说明 |`。\n\n## 输入\n\n见 `schema/review-input.example.json` 与 `schema/README.md`。\n\n## 操作流程\n\n1. 安装本 skill（见仓库根 README：`npx skills add` / ClawHub / clone）。 \n2. Agent 采集会话/用量，**Glob/ls 核对磁盘**后再归因。 \n3. 填 `review-input`（或对话等价结构）；需要图时先出 PNG 再填 `dashboard.charts`。 \n4. 按六章输出，或 `python -m cli --input review-input.json -o YYYY-MM-DD周度复盘.md`。 \n5. 对开放会话当场拍板。定时场景见仓库 `automations/`。\n\n### CLI\n\n```bash\ncd {SKILL_DIR}\npython -m cli --input schema/review-input.example.json -o 周度复盘.md\n# 图表默认写到报告同目录；关闭：加 --no-charts\n```\n\n### MCP\n\n工具 `run_weekly_review`，参数 `review_input`（对象）。"},{"path":"legacy/README.md","content":"# Legacy（非公开主路径）\n\n此处保留旧版「直接读 SQLite 会话库」分析器，仅供对照或迁移。\n\n**公开能力请使用：**\n\n1. Agent 按 `SKILL.md` 采集本平台数据 \n2. 填入 `schema/review-input.example.json` 结构 \n3. `python -m cli --input review-input.json -o report.md`"},{"path":"README.md","content":"# AI 用量与提示词复盘助手（weekly-review）\n\n面向任意 AI Agent 的自动复盘 Skill。\n\n## 解决什么问题\n\n1. 提升提示词表达能力 \n2. 查看本周 AI 用量 \n3. 总结工作低效与高效 \n4. 复盘可视化图表（时间分布 / 归因 / Token） \n5. 清理/对齐开放对话 \n6. AI 自动/定时周复盘 \n\n## Install\n\n```bash\nnpx skills add testman2025/weekly-review-skill --skill weekly-review\n# or\nopenclaw skills install weekly-review\n```\n\nSee the repo root [README](../../README.md).\n\n## Output shape (locked)\n\n1. Dashboard table (`指标 | 数值 | 口径说明`) + one-liner + optional chart paths \n2. Theme project sections with dimension tables (incl. disk verification) \n3. Problems + three-class root cause + positives \n4. Action ledger \n5. Open sessions \n6. Automations overview \n7. Fact corrections (optional)\n\n## Render\n\n```bash\npython -m cli --input schema/review-input.example.json -o weekly-report.md\n```\n\nSee `SKILL.md` and `schema/`."},{"path":"schema/README.md","content":"# review-input 约定\n\n对齐定稿模板：`# YYYY-MM-DD 周度复盘（weekly-review skill · 六章结构）`。\n\nAgent 采集事实后填本结构；skill 只渲染结构。**图表**用 `dashboard.charts` 引用已生成的 PNG 路径（辅助图表），不以 Mermaid 为主。\n\n## 图表\n\n渲染默认生成两张**定稿样式 SVG**（标准库）：\n\n1. `chart-时间分布-*.svg` — 横向条形「本周时间投入分布」\n2. `chart-归因Token-*.svg` — 左右双柱「低效条目归因」+「Token 消耗趋势」\n\n数据优先读 `charts_data`：\n\n```json\n\"charts_data\": {\n \"time_distribution\": [{ \"label\": \"社媒 Agent 主题\", \"hours\": 15.4 }],\n \"attribution\": [{ \"label\": \"【用户】指令模糊\", \"count\": 3 }],\n \"token_trend\": [{ \"label\": \"本周\", \"millions\": 2.74 }]\n}\n```\n\n报告用 `![...](path)` 嵌入。`--no-charts` 可关闭。\n\n\n| 字段 | 说明 |\n|------|------|\n| `period` | `start` / `end`（必填） |\n| `meta` | `title` / `methodology` / `version_note` / `task_split` |\n| `source.note` | 数据来源一行说明 |\n| `dashboard.metrics[]` | `{name,value,note}` 看板表 |\n| `dashboard.summary` | 一句话结论 |\n| `dashboard.charts[]` | 辅助图 PNG 相对路径 |\n| `projects[]` | `{title, dimensions{做了什么,…}}` 主题分节 |\n| `other_projects[]` | `{name,hours,note}` |\n| `problems[]` / `positives[]` | 问题表 + 正面范本 |\n| `actions` | `done` / `observing` / `pending` |\n| `open_sessions[]` | 待对齐会话 |\n| `automations.narrative` | 自动化概览（条目列表或长文） |\n| `corrections*` | 可选第七章事实更正 |\n\n完整示例：`review-input.example.json`。"},{"path":"_meta.json","content":"{\n \"ownerId\": \"kn71svn18t9cb73emms0vf78s984tb7p\",\n \"slug\": \"weekly-review\",\n \"version\": \"1.2.5\",\n \"publishedAt\": 1784863899454\n}"}],"languages":[],"docsSourceLabel":"CLAWHUB","editorialOverview":"面向任意 AI Agent 自动复盘的 Skill（weekly-review / 复盘 / 提示词优化 / 会话清理）。 核心能力覆盖：AI 用量与时间看板、提示词复盘与改写建议、高效/低效归因、 时间分布与 Token/归因趋势图、开放会话对齐清理、定时自动周报。 当用户任务涉及周度复盘、用量查看、提示词改进... Skill: Weekly Review Owner: testman2025 Summary: 面向任意 AI Agent 自动复盘的 Skill（weekly-review / 复盘 / 提示词优化 / 会话清理）。 核心能力覆盖：AI 用量与时间看板、提示词复盘与改写建议、高效/低效归因、 时间分布与 Token/归因趋势图、开放会话对齐清理、定时自动周报。 当用户任务涉及周度复盘、用量查看、提示词改进... Tags: agent-agnostic:1.1.2, agent-fed:1.2.3, automation:1.1.0, latest:1.2.5, office:1.1.0, productivity:1.2.1, retrospective:1.2.3, weekly-review:1.2.3 Version history: v1.2.5 | 2026-07-24T03:31:39.454Z | user","editorialQuality":{"score":100,"threshold":65,"status":"ready","wordCount":731,"uniquenessScore":58,"reasons":[]}},"media":{"evidence":{"source":"no-media","verified":false,"confidence":"low","updatedAt":"2026-10-11T14:45:34.190Z","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":"2026-10-11T14:45:34.190Z","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":"2026-10-11T17:43:31.300Z","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"}]}}}