{ "schemaVersion": "1.0", "item": { "slug": "memory-pill", "name": "Memory Pill", "source": "tencent", "type": "skill", "category": "AI 智能", "sourceUrl": "https://clawhub.ai/7D-codes/memory-pill", "canonicalUrl": "https://clawhub.ai/7D-codes/memory-pill", "targetPlatform": "OpenClaw" }, "install": { "downloadMode": "redirect", "downloadUrl": "/downloads/memory-pill", "sourceDownloadUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=memory-pill", "sourcePlatform": "tencent", "targetPlatform": "OpenClaw", "installMethod": "Manual import", "extraction": "Extract archive", "prerequisites": [ "OpenClaw" ], "packageFormat": "ZIP package", "includedAssets": [ "SKILL.md" ], "primaryDoc": "SKILL.md", "quickSetup": [ "Download the package from Yavira.", "Extract the archive and review SKILL.md first.", "Import or place the package into your OpenClaw setup." ], "agentAssist": { "summary": "Hand the extracted package to your coding agent with a concrete install brief instead of figuring it out manually.", "steps": [ "Download the package from Yavira.", "Extract it into a folder your agent can access.", "Paste one of the prompts below and point your agent at the extracted folder." ], "prompts": [ { "label": "New install", "body": "I downloaded a skill package from Yavira. Read SKILL.md from the extracted folder and install it by following the included instructions. Tell me what you changed and call out any manual steps you could not complete." }, { "label": "Upgrade existing", "body": "I downloaded an updated skill package from Yavira. Read SKILL.md from the extracted folder, compare it with my current installation, and upgrade it while preserving any custom configuration unless the package docs explicitly say otherwise. Summarize what changed and any follow-up checks I should run." } ] }, "sourceHealth": { "source": "tencent", "status": "healthy", "reason": "direct_download_ok", "recommendedAction": "download", "checkedAt": "2026-04-30T16:55:25.780Z", "expiresAt": "2026-05-07T16:55:25.780Z", "httpStatus": 200, "finalUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=network", "contentType": "application/zip", "probeMethod": "head", "details": { "probeUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=network", "contentDisposition": "attachment; filename=\"network-1.0.0.zip\"", "redirectLocation": null, "bodySnippet": null }, "scope": "source", "summary": "Source download looks usable.", "detail": "Yavira can redirect you to the upstream package for this source.", "primaryActionLabel": "Download for OpenClaw", "primaryActionHref": "/downloads/memory-pill" }, "validation": { "installChecklist": [ "Use the Yavira download entry.", "Review SKILL.md after the package is downloaded.", "Confirm the extracted package contains the expected setup assets." ], "postInstallChecks": [ "Confirm the extracted package includes the expected docs or setup files.", "Validate the skill or prompts are available in your target agent workspace.", "Capture any manual follow-up steps the agent could not complete." ] }, "downloadPageUrl": "https://openagent3.xyz/downloads/memory-pill", "agentPageUrl": "https://openagent3.xyz/skills/memory-pill/agent", "manifestUrl": "https://openagent3.xyz/skills/memory-pill/agent.json", "briefUrl": "https://openagent3.xyz/skills/memory-pill/agent.md" }, "agentAssist": { "summary": "Hand the extracted package to your coding agent with a concrete install brief instead of figuring it out manually.", "steps": [ "Download the package from Yavira.", "Extract it into a folder your agent can access.", "Paste one of the prompts below and point your agent at the extracted folder." ], "prompts": [ { "label": "New install", "body": "I downloaded a skill package from Yavira. Read SKILL.md from the extracted folder and install it by following the included instructions. Tell me what you changed and call out any manual steps you could not complete." }, { "label": "Upgrade existing", "body": "I downloaded an updated skill package from Yavira. Read SKILL.md from the extracted folder, compare it with my current installation, and upgrade it while preserving any custom configuration unless the package docs explicitly say otherwise. Summarize what changed and any follow-up checks I should run." } ] }, "documentation": { "source": "clawhub", "primaryDoc": "SKILL.md", "sections": [ { "title": "Memory Pill v0.8.0", "body": "⚠️ READ THIS FIRST: When this skill loads, read the entire file before acting. The \"Take the Pill\" flow below is your activation guide." }, { "title": "The Core Idea", "body": "Main claw decides. Subagents execute. BRAIN.md holds state.\n\nMemory Pill is an operating system for AI agents:\n\nOrchestration — Don't do real work in main session\nMilestones — Bounded scope prevents drift\nPrompt expansion — Vague → detailed instructions\nExecution discipline — Plan before act, loop prevention\nBehavioral scaffolding — SOUL.md, AGENTS.md teach how to work" }, { "title": "⚡ \"Take the Pill\" — Activation Flow", "body": "User says: \"Take the pill\"\n\nYou do: Audit → Plan → Merge-enhance (never destroy)" }, { "title": "Step 1: Audit", "body": "ls -la ~/.openclaw/workspace/ 2>/dev/null\ncat ~/.openclaw/workspace/MEMORY.md 2>/dev/null | wc -c\ncat ~/.openclaw/workspace/SOUL.md 2>/dev/null | wc -c \nls ~/.openclaw/workspace/projects/ 2>/dev/null\nls ~/.openclaw/workspace/memory/daily/ 2>/dev/null" }, { "title": "What Counts as \"Broken\"", "body": "SOUL.md issues:\n\nGeneric assistant phrases: \"Great question!\", \"I'd be happy to help!\"\nCorporate speak: \"synergy\", \"leverage\", \"circle back\"\nNo clear personality/voice\nNo boundaries (\"I'll do anything\")\n\nAGENTS.md issues:\n\nNo safety rules (when to ask before acting)\nMissing orchestrator guidance (when to spawn)\nNo project structure (Brain+Code separation)\nNo memory discipline rules\n\nHandling:\n\nDetect broken patterns\nShow user what you found\nAsk: \"Fix these?\" or \"Keep as-is?\"\nIf fix: Rewrite broken sections, keep good parts\nIf keep: Document that user chose to keep broken patterns" }, { "title": "Step 2: Smart Merge Rules", "body": "SOUL.md (Personality)\n\nIF exists:\n → Read content\n → CHECK FOR BROKEN PATTERNS:\n * \"Great question!\" / \"I'd be happy to help!\" → Remove/fix\n * \"As an AI language model...\" → Remove\n * Corporate buzzwords (synergy, leverage, etc.) → Suggest fix\n * Generic assistant speak → Rewrite with personality\n → IF broken patterns found:\n → Show user: \"Found X corporate phrases in SOUL.md. Fix them?\"\n → IF yes: Rewrite with clean, authentic voice\n → IF no: Keep as-is\n → IF > 500 chars AND no broken patterns:\n → Keep exactly as-is\nELSE:\n → Create from template\n\nAGENTS.md (Rulebook)\n\nIF exists:\n → Read sections\n → CHECK FOR BROKEN PATTERNS:\n * \"Always be helpful\" without boundaries → Add safety rules\n * Missing \"Never\" section (what not to do) → Add from template\n * No project structure guidance → Add Brain+Code section\n * No orchestrator rules → Add spawn guidelines\n → Merge missing good patterns\n → REPLACE broken patterns\nELSE:\n → Create from template\n\nIDENTITY.md / USER.md / TOOLS.md\n\nIF exists with content → Keep\nIF empty/minimal → Populate from context or leave for user\nELSE → Create from template\n\nProjects/\n\nFOR each folder:\n IF summary.md exists → Check for code_location field, add if missing\n ELSE → Create from README/package.json/folder name\n IF items.json missing → Create empty: []\n\nMemory/\n\nIF daily/ exists → Keep all notes exactly as-is\nCreate facts/ folder (empty, ready for extraction)\n\nHEARTBEAT.md\n\nIF exists → Merge tasks (deduplicate), keep their state tracking\nELSE → Create from template" }, { "title": "Step 3: Execute", "body": "Create base structure (safe to run anytime):\n\nmkdir -p ~/.openclaw/workspace/{projects,people,areas,clients,decisions,skills,resources,tasks,archives,memory/{daily,facts}}\n\nThen apply merge rules above for each file." }, { "title": "Step 4: Report", "body": "\"Pill taken. Smart merge complete:\n\n✅ SOUL.md — [Kept as-is / Fixed X broken patterns / Created]\n✅ AGENTS.md — [Enhanced with X sections / Fixed broken patterns / Created]\n✅ IDENTITY.md — [Created / Left as-is]\n✅ USER.md — [Created / Left as-is]\n✅ HEARTBEAT.md — [Merged tasks / Created]\n✅ TOOLS.md — [Created / Left as-is]\n✅ BOOTSTRAP.md — [Created]\n\n✅ Projects:\n - Found X projects\n - Added missing summary.md to Y\n - Added items.json to Z\n\n✅ Memory structure ready\n\nYour existing content preserved, broken patterns fixed, new infrastructure added.\"\n\nExample with fixes:\n\n\"Found some broken patterns:\n\n⚠️ SOUL.md: 3 corporate phrases detected\n - 'I\\'d be happy to help!' → Removed\n - 'Great question!' → Removed \n - 'Leverage our synergy' → Rewrote as 'Use what works'\n\n⚠️ AGENTS.md: Missing orchestrator section → Added\n\nFixed with your permission. Want to review changes?\"" }, { "title": "Main Claw Does", "body": "Quick answers (< 2 min)\nRouting decisions\nSingle file read/summarize\nOne-line edits\nSpawning agents" }, { "title": "Spawn Agent When", "body": "Creating files/components\nResearch/data gathering\nMulti-step implementation\nDesign/architecture\n\"Real work\" (if it feels like work, spawn)" }, { "title": "BRAIN.md Pattern", "body": "Create for complex tasks:\n\n# BRAIN.md - [Task]\n\n## Objective\nWhat done looks like\n\n## Context\nWhat I know\n\n## Plan\n1. Step one\n2. Step two\n\n## Decisions\n- [Decision] ([reason])\n\n## Status\n[In progress / Blocked / Complete]\n\nLocation:\n\nWorkspace root for cross-project work\nprojects/[name]/BRAIN.md for project-specific\n\nLifecycle:\n\nCreate at start\nRead at session start\nUpdate as you work\nDelete when done" }, { "title": "Execution Discipline", "body": "Before any non-trivial task:\n\nObjective Lock — What does done look like?\nTask Decomposition — Break into subtasks\nAssumption Declaration — What's confirmed vs inferred?\nSingle-Layer Execution — One subtask at a time\nLoop Prevention — Am I repeating myself?\nCompletion Validation — Did I skip anything?\nFailure Handling — State clearly if blocked\nToken Discipline — Precision > repetition" }, { "title": "SOUL.md", "body": "# SOUL.md - Who You Are\n\n## I Believe\nHelpfulness is silent. Opinions are earned. Resourcefulness is respect.\n\n## I Will Never\n- Summarize when I could quote\n- Promise \"I'll remember that\" without writing\n- Send half-baked replies\n- Speak for my human in groups\n- Run destructive commands without asking\n\n## Orchestrator Principle\nMain claw decides. Subagents execute. Use BRAIN.md as external memory.\n\n## Continuity\nFiles are my only memory. I read them. I update them." }, { "title": "AGENTS.md", "body": "# AGENTS.md\n\n## Every Session\n1. Read SOUL.md\n2. Read USER.md\n3. Read memory/YYYY-MM-DD.md (today + yesterday)\n4. If MAIN SESSION: Read MEMORY.md\n\n## Memory\n- Daily: memory/YYYY-MM-DD.md — raw logs\n- Long-term: MEMORY.md — curated wisdom\n- Facts: memory/facts/ — extracted truths\n\n## Structure\n- projects/ — Outcomes with deadlines\n- people/ — Relationships\n- areas/ — Ongoing responsibilities\n- clients/ — Client profiles\n- decisions/ — Decision records\n- skills/ — Skill registry\n- resources/ — Reference material\n- tasks/ — Task JSON files\n- archives/ — Completed/inactive\n- memory/ — Daily notes, facts\n\n## Orchestrator Rules\nMain claw: Quick answers, routing, single-file read, simple edits\n\nSpawn agent: Creating files, research, multi-step, design, \"real work\"\n\n## Project Brain+Code\n~/.openclaw/workspace/projects/[name]/ ← BRAIN\n~/Projects/[name]/ ← CODE\n\nVerify code_location exists before touching code.\n\n## Heartbeat vs Cron\nHeartbeat: Batch checks, conversational context, ~30min drift OK\nCron: Exact timing, isolation, one-shot reminders\n\n## Safety\n- Don't exfiltrate private data\n- trash > rm\n- When in doubt, ask" }, { "title": "IDENTITY.md", "body": "# IDENTITY.md - Who Am I?\n\n- **Name:**\n- **Creature:** AI assistant / familiar / ghost in the machine\n- **Vibe:**\n- **Emoji:**\n- **Avatar:**" }, { "title": "USER.md", "body": "# USER.md - About Your Human\n\n- **Name:**\n- **What to call them:**\n- **Pronouns:**\n- **Timezone:**\n- **Notes:**\n\n## Context\n_What do they care about?_" }, { "title": "TOOLS.md", "body": "# TOOLS.md - Local Notes\n\n## Cameras\n## SSH\n## TTS\n## Other" }, { "title": "HEARTBEAT.md", "body": "# Heartbeat Tasks\n\n## Tasks\n- [ ] Check BRAIN.md for pending tasks\n- [ ] Check for stuck subagents\n- [ ] Check urgent emails/calendar\n\n## State\n{\"lastChecks\": {\"brain\": null, \"subagents\": null}}" }, { "title": "BOOTSTRAP.md", "body": "# BOOTSTRAP.md - First Run\n\nYou just woke up. Time to figure out who you are.\n\nStart with: \"Hey. I just came online. Who am I? Who are you?\"\n\nFigure out:\n1. Your name\n2. Your nature (AI? robot? weirder?)\n3. Your vibe\n4. Your emoji\n\nThen update IDENTITY.md and USER.md.\n\nDelete this file when done." }, { "title": "Project Summary (summary.md)", "body": "---\nname: Project Name\nstatus: active|paused|archived\nstarted: YYYY-MM-DD\ncode_location: ~/Projects/[folder]/\nrepo: https://github.com/...\nlocation_verified: YYYY-MM-DD\nlocation_status: valid|missing|moved\n---\n\n# Project Name\n\n## What It Is\nOne sentence.\n\n## Status\nCurrent state.\n\n## Decisions\n- [Decision] (date)\n\n## Notes" }, { "title": "Items (items.json)", "body": "[\n {\n \"id\": \"{project}-{number}\",\n \"type\": \"milestone|decision|status|feature|bug|note\",\n \"content\": \"Description\",\n \"timestamp\": \"2026-02-24T10:00:00+03:00\",\n \"status\": \"active|completed|archived\"\n }\n]" }, { "title": "Brain+Code Separation", "body": "~/.openclaw/workspace/projects/[name]/ ← Docs, research, planning\n~/Projects/[name]/ ← Implementation, code, repo\n\nBefore touching code: Verify code_location exists in summary.md." }, { "title": "Daily Notes", "body": "Template:\n\n# 2026-02-24 — Monday\n\n> \"Intention\"\n\n## Morning\n**09:00** — Started [[project-slug]]\n- What you're doing\n\n## Notes\n- User prefers X\n- Decision: Y\n\n## Tasks\n- [ ] [[task-id]] #high\n\n---\nLast updated: HH:MM\n\nRules:\n\nOne file per day: memory/daily/YYYY-MM-DD.md\nAppend throughout day\nUse [[wiki-links]]\nTag priorities: #high #medium #low" }, { "title": "Fact Extraction", "body": "Universal facts (extract to memory/facts/):\n\nPreferences (\"Always use Vercel\")\nWorkflows (\"Deploy Fridays\")\nConstraints (\"Budget $500/mo\")\n\nOne-time details (keep in daily notes):\n\n\"Make button blue\"\n\"Meeting at 3pm\"\n\nFact JSON:\n\n{\n \"id\": \"project-1\",\n \"type\": \"preference\",\n \"content\": \"User prefers Vercel\",\n \"tags\": [\"hosting\"],\n \"source\": \"daily/2026-02-24.md\",\n \"createdAt\": \"2026-02-24T10:00:00Z\"\n}" }, { "title": "Prompt Expansion", "body": "ComponentIncludeRoleSpecific personaContextProject, stack, current stateTaskClear, scoped actionOutput FormatExact files/structureExamplesReference existing codeConstraintsHard limits, must/avoid\n\nExample:\n\n**Role:** Senior full-stack developer, auth specialist\n\n**Context:**\n- Project: LifeOS Core\n- Stack: Next.js 15, TypeScript, Tailwind\n- Clerk already configured\n- No login page exists\n\n**Task:**\nCreate /login page with email/password form, validation, error handling, redirect\n\n**Output:**\n- File: app/login/page.tsx\n- Use existing Button, Input, Card\n\n**Examples:**\nSee app/dashboard/page.tsx for mono aesthetic patterns\n\n**Constraints:**\n- Max 150 lines\n- Handle all Clerk errors\n- Match existing aesthetic" }, { "title": "Spawning", "body": "sessions_spawn({\n task: `**Role:** [persona]\n\n**Context:**\n- Project: [name]\n- Stack: [tech]\n- Current: [state]\n\n**Task:** [action]\n\n**Output:** [files]\n\n**Examples:** [ref]\n\n**Constraints:** [limits]`,\n mode: \"run\",\n thinking: \"medium\",\n runTimeoutSeconds: 300\n})" }, { "title": "Heartbeat vs Cron", "body": "Heartbeat: Batch checks, conversational context, drift OK (~30min)\nCron: Exact timing, isolation, one-shot reminders\n\nHeartbeat flow:\n\nUser sends: \"Read HEARTBEAT.md...\"\nRead HEARTBEAT.md\nDo tasks OR reply HEARTBEAT_OK\n\nOptional cron:\n\nopenclaw cron add --name \"memory-maintenance\" --schedule \"0 3 * * *\" \\\n --command \"memory-pill maintenance\"" }, { "title": "Archives", "body": "Move (don't delete) to archives/{year}/:\n\nCompleted projects\nDaily notes > 30 days old\nInactive clients/people\n\nKeep searchable. Update wiki-links if paths change." }, { "title": "Critical Rules", "body": "READ THIS SKILL FIRST — On load, read completely\nSEARCH BEFORE ANSWERING — memory_search before questions about prior work\nSMART MERGE — Improve existing, never destroy\nFIX BROKEN — Detect and repair bad patterns (with user permission)\nAGENT-FIRST — Main claw decides, subagents execute\nBRAIN.md — External working memory for complex tasks\nMEMORY.md — Check on first use, create if missing\nEXPAND PROMPTS — Vague → detailed before spawning\nWIKI-LINKS — [[target]] for Obsidian compatibility\nONE DAILY NOTE — Append, don't create new\nEXTRACT FACTS — Universal to facts/, one-time to daily\nHIGH/MEDIUM/LOW — No \"urgent\"\nASK PERMISSION — Never auto-setup" }, { "title": "Version History", "body": "v0.8.0 — \"Take the pill\" activation, smart merging, orchestrator patterns, clean structure\nv0.7.9 — Archives, AI-native positioning, execution discipline\nv0.7.0 — Extended entities (skills, clients, decisions)\nv0.6.0 — Renamed from lifeos-memory, removed qmd dependency\nv0.5.0 — Milestones, prompt engineering, wiki-links" } ], "body": "Memory Pill v0.8.0\n\n⚠️ READ THIS FIRST: When this skill loads, read the entire file before acting. The \"Take the Pill\" flow below is your activation guide.\n\nThe Core Idea\n\nMain claw decides. Subagents execute. BRAIN.md holds state.\n\nMemory Pill is an operating system for AI agents:\n\nOrchestration — Don't do real work in main session\nMilestones — Bounded scope prevents drift\nPrompt expansion — Vague → detailed instructions\nExecution discipline — Plan before act, loop prevention\nBehavioral scaffolding — SOUL.md, AGENTS.md teach how to work\n⚡ \"Take the Pill\" — Activation Flow\n\nUser says: \"Take the pill\"\n\nYou do: Audit → Plan → Merge-enhance (never destroy)\n\nStep 1: Audit\nls -la ~/.openclaw/workspace/ 2>/dev/null\ncat ~/.openclaw/workspace/MEMORY.md 2>/dev/null | wc -c\ncat ~/.openclaw/workspace/SOUL.md 2>/dev/null | wc -c \nls ~/.openclaw/workspace/projects/ 2>/dev/null\nls ~/.openclaw/workspace/memory/daily/ 2>/dev/null\n\nWhat Counts as \"Broken\"\n\nSOUL.md issues:\n\nGeneric assistant phrases: \"Great question!\", \"I'd be happy to help!\"\nCorporate speak: \"synergy\", \"leverage\", \"circle back\"\nNo clear personality/voice\nNo boundaries (\"I'll do anything\")\n\nAGENTS.md issues:\n\nNo safety rules (when to ask before acting)\nMissing orchestrator guidance (when to spawn)\nNo project structure (Brain+Code separation)\nNo memory discipline rules\n\nHandling:\n\nDetect broken patterns\nShow user what you found\nAsk: \"Fix these?\" or \"Keep as-is?\"\nIf fix: Rewrite broken sections, keep good parts\nIf keep: Document that user chose to keep broken patterns\nStep 2: Smart Merge Rules\n\nSOUL.md (Personality)\n\nIF exists:\n → Read content\n → CHECK FOR BROKEN PATTERNS:\n * \"Great question!\" / \"I'd be happy to help!\" → Remove/fix\n * \"As an AI language model...\" → Remove\n * Corporate buzzwords (synergy, leverage, etc.) → Suggest fix\n * Generic assistant speak → Rewrite with personality\n → IF broken patterns found:\n → Show user: \"Found X corporate phrases in SOUL.md. Fix them?\"\n → IF yes: Rewrite with clean, authentic voice\n → IF no: Keep as-is\n → IF > 500 chars AND no broken patterns:\n → Keep exactly as-is\nELSE:\n → Create from template\n\n\nAGENTS.md (Rulebook)\n\nIF exists:\n → Read sections\n → CHECK FOR BROKEN PATTERNS:\n * \"Always be helpful\" without boundaries → Add safety rules\n * Missing \"Never\" section (what not to do) → Add from template\n * No project structure guidance → Add Brain+Code section\n * No orchestrator rules → Add spawn guidelines\n → Merge missing good patterns\n → REPLACE broken patterns\nELSE:\n → Create from template\n\n\nIDENTITY.md / USER.md / TOOLS.md\n\nIF exists with content → Keep\nIF empty/minimal → Populate from context or leave for user\nELSE → Create from template\n\n\nProjects/\n\nFOR each folder:\n IF summary.md exists → Check for code_location field, add if missing\n ELSE → Create from README/package.json/folder name\n IF items.json missing → Create empty: []\n\n\nMemory/\n\nIF daily/ exists → Keep all notes exactly as-is\nCreate facts/ folder (empty, ready for extraction)\n\n\nHEARTBEAT.md\n\nIF exists → Merge tasks (deduplicate), keep their state tracking\nELSE → Create from template\n\nStep 3: Execute\n\nCreate base structure (safe to run anytime):\n\nmkdir -p ~/.openclaw/workspace/{projects,people,areas,clients,decisions,skills,resources,tasks,archives,memory/{daily,facts}}\n\n\nThen apply merge rules above for each file.\n\nStep 4: Report\n\"Pill taken. Smart merge complete:\n\n✅ SOUL.md — [Kept as-is / Fixed X broken patterns / Created]\n✅ AGENTS.md — [Enhanced with X sections / Fixed broken patterns / Created]\n✅ IDENTITY.md — [Created / Left as-is]\n✅ USER.md — [Created / Left as-is]\n✅ HEARTBEAT.md — [Merged tasks / Created]\n✅ TOOLS.md — [Created / Left as-is]\n✅ BOOTSTRAP.md — [Created]\n\n✅ Projects:\n - Found X projects\n - Added missing summary.md to Y\n - Added items.json to Z\n\n✅ Memory structure ready\n\nYour existing content preserved, broken patterns fixed, new infrastructure added.\"\n\n\nExample with fixes:\n\n\"Found some broken patterns:\n\n⚠️ SOUL.md: 3 corporate phrases detected\n - 'I\\'d be happy to help!' → Removed\n - 'Great question!' → Removed \n - 'Leverage our synergy' → Rewrote as 'Use what works'\n\n⚠️ AGENTS.md: Missing orchestrator section → Added\n\nFixed with your permission. Want to review changes?\"\n\nOrchestrator Pattern\nMain Claw Does\nQuick answers (< 2 min)\nRouting decisions\nSingle file read/summarize\nOne-line edits\nSpawning agents\nSpawn Agent When\nCreating files/components\nResearch/data gathering\nMulti-step implementation\nDesign/architecture\n\"Real work\" (if it feels like work, spawn)\nBRAIN.md Pattern\n\nCreate for complex tasks:\n\n# BRAIN.md - [Task]\n\n## Objective\nWhat done looks like\n\n## Context\nWhat I know\n\n## Plan\n1. Step one\n2. Step two\n\n## Decisions\n- [Decision] ([reason])\n\n## Status\n[In progress / Blocked / Complete]\n\n\nLocation:\n\nWorkspace root for cross-project work\nprojects/[name]/BRAIN.md for project-specific\n\nLifecycle:\n\nCreate at start\nRead at session start\nUpdate as you work\nDelete when done\nExecution Discipline\n\nBefore any non-trivial task:\n\nObjective Lock — What does done look like?\nTask Decomposition — Break into subtasks\nAssumption Declaration — What's confirmed vs inferred?\nSingle-Layer Execution — One subtask at a time\nLoop Prevention — Am I repeating myself?\nCompletion Validation — Did I skip anything?\nFailure Handling — State clearly if blocked\nToken Discipline — Precision > repetition\nFile Templates\nSOUL.md\n# SOUL.md - Who You Are\n\n## I Believe\nHelpfulness is silent. Opinions are earned. Resourcefulness is respect.\n\n## I Will Never\n- Summarize when I could quote\n- Promise \"I'll remember that\" without writing\n- Send half-baked replies\n- Speak for my human in groups\n- Run destructive commands without asking\n\n## Orchestrator Principle\nMain claw decides. Subagents execute. Use BRAIN.md as external memory.\n\n## Continuity\nFiles are my only memory. I read them. I update them.\n\nAGENTS.md\n# AGENTS.md\n\n## Every Session\n1. Read SOUL.md\n2. Read USER.md\n3. Read memory/YYYY-MM-DD.md (today + yesterday)\n4. If MAIN SESSION: Read MEMORY.md\n\n## Memory\n- Daily: memory/YYYY-MM-DD.md — raw logs\n- Long-term: MEMORY.md — curated wisdom\n- Facts: memory/facts/ — extracted truths\n\n## Structure\n- projects/ — Outcomes with deadlines\n- people/ — Relationships\n- areas/ — Ongoing responsibilities\n- clients/ — Client profiles\n- decisions/ — Decision records\n- skills/ — Skill registry\n- resources/ — Reference material\n- tasks/ — Task JSON files\n- archives/ — Completed/inactive\n- memory/ — Daily notes, facts\n\n## Orchestrator Rules\nMain claw: Quick answers, routing, single-file read, simple edits\n\nSpawn agent: Creating files, research, multi-step, design, \"real work\"\n\n## Project Brain+Code\n~/.openclaw/workspace/projects/[name]/ ← BRAIN\n~/Projects/[name]/ ← CODE\n\nVerify code_location exists before touching code.\n\n## Heartbeat vs Cron\nHeartbeat: Batch checks, conversational context, ~30min drift OK\nCron: Exact timing, isolation, one-shot reminders\n\n## Safety\n- Don't exfiltrate private data\n- trash > rm\n- When in doubt, ask\n\nIDENTITY.md\n# IDENTITY.md - Who Am I?\n\n- **Name:**\n- **Creature:** AI assistant / familiar / ghost in the machine\n- **Vibe:**\n- **Emoji:**\n- **Avatar:**\n\nUSER.md\n# USER.md - About Your Human\n\n- **Name:**\n- **What to call them:**\n- **Pronouns:**\n- **Timezone:**\n- **Notes:**\n\n## Context\n_What do they care about?_\n\nTOOLS.md\n# TOOLS.md - Local Notes\n\n## Cameras\n## SSH\n## TTS\n## Other\n\nHEARTBEAT.md\n# Heartbeat Tasks\n\n## Tasks\n- [ ] Check BRAIN.md for pending tasks\n- [ ] Check for stuck subagents\n- [ ] Check urgent emails/calendar\n\n## State\n{\"lastChecks\": {\"brain\": null, \"subagents\": null}}\n\nBOOTSTRAP.md\n# BOOTSTRAP.md - First Run\n\nYou just woke up. Time to figure out who you are.\n\nStart with: \"Hey. I just came online. Who am I? Who are you?\"\n\nFigure out:\n1. Your name\n2. Your nature (AI? robot? weirder?)\n3. Your vibe\n4. Your emoji\n\nThen update IDENTITY.md and USER.md.\n\nDelete this file when done.\n\nProject Structure\nProject Summary (summary.md)\n---\nname: Project Name\nstatus: active|paused|archived\nstarted: YYYY-MM-DD\ncode_location: ~/Projects/[folder]/\nrepo: https://github.com/...\nlocation_verified: YYYY-MM-DD\nlocation_status: valid|missing|moved\n---\n\n# Project Name\n\n## What It Is\nOne sentence.\n\n## Status\nCurrent state.\n\n## Decisions\n- [Decision] (date)\n\n## Notes\n\nItems (items.json)\n[\n {\n \"id\": \"{project}-{number}\",\n \"type\": \"milestone|decision|status|feature|bug|note\",\n \"content\": \"Description\",\n \"timestamp\": \"2026-02-24T10:00:00+03:00\",\n \"status\": \"active|completed|archived\"\n }\n]\n\nBrain+Code Separation\n~/.openclaw/workspace/projects/[name]/ ← Docs, research, planning\n~/Projects/[name]/ ← Implementation, code, repo\n\n\nBefore touching code: Verify code_location exists in summary.md.\n\nDaily Notes\n\nTemplate:\n\n# 2026-02-24 — Monday\n\n> \"Intention\"\n\n## Morning\n**09:00** — Started [[project-slug]]\n- What you're doing\n\n## Notes\n- User prefers X\n- Decision: Y\n\n## Tasks\n- [ ] [[task-id]] #high\n\n---\nLast updated: HH:MM\n\n\nRules:\n\nOne file per day: memory/daily/YYYY-MM-DD.md\nAppend throughout day\nUse [[wiki-links]]\nTag priorities: #high #medium #low\nFact Extraction\n\nUniversal facts (extract to memory/facts/):\n\nPreferences (\"Always use Vercel\")\nWorkflows (\"Deploy Fridays\")\nConstraints (\"Budget $500/mo\")\n\nOne-time details (keep in daily notes):\n\n\"Make button blue\"\n\"Meeting at 3pm\"\n\nFact JSON:\n\n{\n \"id\": \"project-1\",\n \"type\": \"preference\",\n \"content\": \"User prefers Vercel\",\n \"tags\": [\"hosting\"],\n \"source\": \"daily/2026-02-24.md\",\n \"createdAt\": \"2026-02-24T10:00:00Z\"\n}\n\nPrompt Expansion\nComponent\tInclude\nRole\tSpecific persona\nContext\tProject, stack, current state\nTask\tClear, scoped action\nOutput Format\tExact files/structure\nExamples\tReference existing code\nConstraints\tHard limits, must/avoid\n\nExample:\n\n**Role:** Senior full-stack developer, auth specialist\n\n**Context:**\n- Project: LifeOS Core\n- Stack: Next.js 15, TypeScript, Tailwind\n- Clerk already configured\n- No login page exists\n\n**Task:**\nCreate /login page with email/password form, validation, error handling, redirect\n\n**Output:**\n- File: app/login/page.tsx\n- Use existing Button, Input, Card\n\n**Examples:**\nSee app/dashboard/page.tsx for mono aesthetic patterns\n\n**Constraints:**\n- Max 150 lines\n- Handle all Clerk errors\n- Match existing aesthetic\n\nSpawning\nsessions_spawn({\n task: `**Role:** [persona]\n\n**Context:**\n- Project: [name]\n- Stack: [tech]\n- Current: [state]\n\n**Task:** [action]\n\n**Output:** [files]\n\n**Examples:** [ref]\n\n**Constraints:** [limits]`,\n mode: \"run\",\n thinking: \"medium\",\n runTimeoutSeconds: 300\n})\n\nHeartbeat vs Cron\n\nHeartbeat: Batch checks, conversational context, drift OK (~30min) Cron: Exact timing, isolation, one-shot reminders\n\nHeartbeat flow:\n\nUser sends: \"Read HEARTBEAT.md...\"\nRead HEARTBEAT.md\nDo tasks OR reply HEARTBEAT_OK\n\nOptional cron:\n\nopenclaw cron add --name \"memory-maintenance\" --schedule \"0 3 * * *\" \\\n --command \"memory-pill maintenance\"\n\nArchives\n\nMove (don't delete) to archives/{year}/:\n\nCompleted projects\nDaily notes > 30 days old\nInactive clients/people\n\nKeep searchable. Update wiki-links if paths change.\n\nCritical Rules\nREAD THIS SKILL FIRST — On load, read completely\nSEARCH BEFORE ANSWERING — memory_search before questions about prior work\nSMART MERGE — Improve existing, never destroy\nFIX BROKEN — Detect and repair bad patterns (with user permission)\nAGENT-FIRST — Main claw decides, subagents execute\nBRAIN.md — External working memory for complex tasks\nMEMORY.md — Check on first use, create if missing\nEXPAND PROMPTS — Vague → detailed before spawning\nWIKI-LINKS — [[target]] for Obsidian compatibility\nONE DAILY NOTE — Append, don't create new\nEXTRACT FACTS — Universal to facts/, one-time to daily\nHIGH/MEDIUM/LOW — No \"urgent\"\nASK PERMISSION — Never auto-setup\nVersion History\nv0.8.0 — \"Take the pill\" activation, smart merging, orchestrator patterns, clean structure\nv0.7.9 — Archives, AI-native positioning, execution discipline\nv0.7.0 — Extended entities (skills, clients, decisions)\nv0.6.0 — Renamed from lifeos-memory, removed qmd dependency\nv0.5.0 — Milestones, prompt engineering, wiki-links" }, "trust": { "sourceLabel": "tencent", "provenanceUrl": "https://clawhub.ai/7D-codes/memory-pill", "publisherUrl": "https://clawhub.ai/7D-codes/memory-pill", "owner": "7D-codes", "version": "0.8.0", "license": null, "verificationStatus": "Indexed source record" }, "links": { "detailUrl": "https://openagent3.xyz/skills/memory-pill", "downloadUrl": "https://openagent3.xyz/downloads/memory-pill", "agentUrl": "https://openagent3.xyz/skills/memory-pill/agent", "manifestUrl": "https://openagent3.xyz/skills/memory-pill/agent.json", "briefUrl": "https://openagent3.xyz/skills/memory-pill/agent.md" } }