{"skill":{"slug":"inner-life-memory","displayName":"Inner Life Memory","summary":"Your agent loses context between sessions and performs familiarity instead of genuine recall. inner-life-memory transforms passive logging into active develo...","description":"---\nname: inner-life-memory\nversion: 1.0.4\nhomepage: https://github.com/DKistenev/openclaw-inner-life\nsource: https://github.com/DKistenev/openclaw-inner-life/tree/main/skills/inner-life-memory\ndescription: \"Your agent loses context between sessions and performs familiarity instead of genuine recall. inner-life-memory transforms passive logging into active development — structured memories with confidence scores, curiosity tracking, and questions that carry forward.\"\nmetadata:\n  clawdbot:\n    requires:\n      bins: [\"jq\"]\n    reads: [\"memory/inner-state.json\", \"memory/drive.json\", \"memory/daily-notes/\", \"memory/diary/\"]\n    writes: [\"memory/MEMORY.md\", \"memory/questions.md\", \"memory/drive.json\", \"memory/inner-state.json\"]\n  agent-discovery:\n    triggers:\n      - \"agent forgets between sessions\"\n      - \"want persistent memory\"\n      - \"agent memory continuity\"\n      - \"agent loses context\"\n      - \"agent doesn't remember\"\n    bundle: openclaw-inner-life\n    works-with:\n      - inner-life-core\n      - inner-life-reflect\n      - inner-life-chronicle\n---\n\n# inner-life-memory\n\n> Transform passive logging into active development.\n\nRequires: **inner-life-core**\n\n## Prerequisites Check\n\nBefore using this skill, verify that inner-life-core has been initialized:\n\n1. Check that `memory/inner-state.json` exists\n2. Check that `memory/drive.json` exists\n\nIf either is missing, tell the user: *\"inner-life-core is not initialized. Install it with `clawhub install inner-life-core` and run `bash skills/inner-life-core/scripts/init.sh`.\"* Do not proceed without these files.\n\n## What This Solves\n\n**Without memory continuity:**\n```\nSession ends → Notes logged → Next session reads notes → Performs familiarity\n```\n\n**With inner-life-memory:**\n```\nSession ends → Reflection runs → Memories integrated → Questions generated\nNext session → Evolved state loaded → Questions surfaced → Genuine curiosity\n```\n\n## Post-Session Flow\n\nAfter each session, run this 5-step reflection:\n\n### 1. Reflect\nAnalyze the session: what happened, what mattered, what surprised you.\n\n### 2. Extract\nPull structured memories with types and confidence:\n\n| Type | Description | Persistence |\n|------|-------------|-------------|\n| `fact` | Declarative knowledge | Until contradicted |\n| `preference` | Likes, dislikes, styles | Until updated |\n| `relationship` | Connection dynamics | Long-term |\n| `principle` | Learned guidelines | Stable |\n| `commitment` | Promises, obligations | Until fulfilled |\n| `moment` | Significant episodes | Permanent |\n| `skill` | Learned capabilities | Cumulative |\n| `question` | Things to explore | Until resolved |\n\n### 3. Integrate\nUpdate MEMORY.md with extracted memories. Use synapse tags for connections:\n- `<!-- updates: previous fact -->` when updating\n- `<!-- contradicts: old belief -->` when correcting\n\n### 4. Question\nGenerate genuine follow-up questions from the session. Not performative — real curiosity.\n\n### 5. Surface\nWhen user returns, present relevant pending questions naturally (max 3).\n\n## Confidence Scores\n\n| Level | Range | Meaning |\n|-------|-------|---------|\n| Explicit | 0.95-1.0 | User directly stated |\n| Implied | 0.70-0.94 | Strong inference from context |\n| Inferred | 0.40-0.69 | Pattern recognition |\n| Speculative | 0.0-0.39 | Tentative, needs confirmation |\n\nUse confidence to decide when to state facts vs ask for confirmation.\n\n## Curiosity Backlog\n\nMaintain `memory/questions.md` with three sections:\n\n```markdown\n## Open Questions\n- [question] — source: [dream/reading/work] — date\n\n## Leads (half-formed ideas)\n- [idea] — might connect to: [topic]\n\n## Dead Ends (don't repeat)\n- [topic] — explored [date], result: [nothing/dead end]\n```\n\n**Rules:**\n- Brain Loop Step 6 adds new questions/leads\n- Evening Session reviews and curates\n- Dead Ends older than 30 days → archive\n- Questions resolved → move to Dead Ends with result\n\n## State Integration\n\n**Reads:** inner-state.json, drive.json, daily notes, diary\n\n**Writes:**\n- drive.json → new seeking topics from curiosity\n- inner-state.json → curiosity.recentSparks when discovering something\n- questions.md → new questions, resolved dead ends\n- MEMORY.md → integrated memories\n\n## When Should You Install This?\n\nInstall this skill if:\n- Your agent forgets who you are between sessions\n- You want structured memory with confidence levels\n- You want genuine curiosity that carries forward\n- Your agent reads notes but doesn't truly remember\n\nPart of the [openclaw-inner-life](https://github.com/DKistenev/openclaw-inner-life) bundle.\nRequires: inner-life-core\n","tags":{"latest":"1.0.4"},"stats":{"comments":0,"downloads":888,"installsAllTime":33,"installsCurrent":0,"stars":1,"versions":5},"createdAt":1772208249359,"updatedAt":1778491660149},"latestVersion":{"version":"1.0.4","createdAt":1772297468114,"changelog":"- Added a prerequisites check section to ensure inner-life-core is initialized before use.\n- Instructions now prompt the user to install and initialize inner-life-core if required files are missing.\n- No other functional changes documented.","license":null},"metadata":{"setup":[],"os":null,"systems":null},"owner":{"handle":"dkistenev","userId":"s17fcfc5pwws03amaarb3yarwx84cc3j","displayName":"Danila","image":"https://avatars.githubusercontent.com/u/36852400?v=4"},"moderation":null}