# Send 2nd Brain to your agent Hand the extracted package to your coding agent with a concrete install brief instead of figuring it out manually. ## Fast path - 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. ## Suggested prompts ### New install ```text 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. ``` ### Upgrade existing ```text 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. ``` ## Machine-readable fields ```json { "schemaVersion": "1.0", "item": { "slug": "2nd-brain", "name": "2nd Brain", "source": "tencent", "type": "skill", "category": "其他", "sourceUrl": "https://clawhub.ai/coderaven/2nd-brain", "canonicalUrl": "https://clawhub.ai/coderaven/2nd-brain", "targetPlatform": "OpenClaw" }, "install": { "downloadUrl": "/downloads/2nd-brain", "sourceDownloadUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=2nd-brain", "sourcePlatform": "tencent", "targetPlatform": "OpenClaw", "packageFormat": "ZIP package", "primaryDoc": "SKILL.md", "includedAssets": [ "SKILL.md", "templates/attachments-index.md", "templates/event.md", "templates/game.md", "templates/idea.md", "templates/media.md" ], "downloadMode": "redirect", "sourceHealth": { "source": "tencent", "slug": "2nd-brain", "status": "healthy", "reason": "direct_download_ok", "recommendedAction": "download", "checkedAt": "2026-04-22T14:04:09.960Z", "expiresAt": "2026-04-29T14:04:09.960Z", "httpStatus": 200, "finalUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=2nd-brain", "contentType": "application/zip", "probeMethod": "head", "details": { "probeUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=2nd-brain", "contentDisposition": "attachment; filename=\"2nd-brain-1.3.0.zip\"", "redirectLocation": null, "bodySnippet": null, "slug": "2nd-brain" }, "scope": "item", "summary": "Item download looks usable.", "detail": "Yavira can redirect you to the upstream package for this item.", "primaryActionLabel": "Download for OpenClaw", "primaryActionHref": "/downloads/2nd-brain" }, "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." ] } }, "links": { "detailUrl": "https://openagent3.xyz/skills/2nd-brain", "downloadUrl": "https://openagent3.xyz/downloads/2nd-brain", "agentUrl": "https://openagent3.xyz/skills/2nd-brain/agent", "manifestUrl": "https://openagent3.xyz/skills/2nd-brain/agent.json", "briefUrl": "https://openagent3.xyz/skills/2nd-brain/agent.md" } } ``` ## Documentation ### Brain Skill — 2nd Brain Knowledge Base A personal knowledge management system for capturing and retrieving information about people, places, things, and ideas. ### When to Use This Skill Brain takes precedence over daily logs for named entities. Trigger this skill when: User asks you to remember someone, something, or somewhere User shares information about a person, place, game, tech, event, media, idea, or organization User expresses a preference about an entity ("I like X at Y restaurant" → update Y's file) User asks about something that might be in the brain ("Who was that guy from...", "What did I think about...") User updates existing knowledge ("Actually, he's 27 now", "I finished that game") Keywords that trigger: "remember", "note that", "met this person", "visited", "played", "watched", "read", "idea:", "what do I know about", "who is", "where was" ⚠️ Do NOT put brain-eligible content in daily logs. If it's a named entity (person, place, restaurant, product, game, etc.), it belongs in brain/, not memory/YYYY-MM-DD.md. Daily logs are for session context and ephemeral notes only. 🚨 MEDIA FILES MUST BE SAVED. When user sends photos/audio/video/PDFs about a brain entry, you MUST save the actual file to attachments/. Transcribing content is NOT the same as saving the file. Do BOTH. ### Data Location All brain data lives in: ~/.openclaw/workspace/brain/ brain/ people/ # Contacts, people you've met places/ # Restaurants, landmarks, venues games/ # Video games and interactions tech/ # Devices, products, specs, gotchas events/ # Conferences, meetups, gatherings media/ # Books, shows, films, podcasts ideas/ # Business ideas, concepts, thoughts orgs/ # Companies, communities, groups ### Search & Retrieval This skill uses OpenClaw's built-in memory_search and memory_get tools, which work out of the box with any configured memory backend. ### Searching Use memory_search for all brain lookups: memory_search("Raven Duran") # find a person memory_search("Mamou Prime restaurant") # find a place memory_search("what games has Raven played") # natural language memory_search works transparently whether the backend is the built-in SQLite indexer or QMD. No direct CLI calls needed. ### Reading a File Use memory_get to read a specific brain file once you know its path: memory_get("brain/people/raven-duran.md") memory_get("brain/places/mamou-prime-sm-podium/mamou-prime-sm-podium.md") ### Direct CLI (Optional / Advanced Only) Only use the qmd CLI directly when searching a non-workspace collection (e.g., the skills collection). For all brain lookups, use memory_search. # Only for skills collection or non-workspace paths: export PATH="$HOME/.bun/bin:$PATH" qmd search "keyword" -c skills ### Creating a New Entry Search first — Run memory_search("") to check for existing entries No match — Create new file using the appropriate template from skills/brain/templates/ Possible clash — List all potential matches and ask user to confirm before creating ### Updating an Existing Entry Find the file — Use memory_search or direct path if known Surgical edit — Update only the relevant section, don't rewrite the whole file Log the date — Add timestamp to Notes or Interactions section Update frontmatter — Bump last_updated field ### Searching / Retrieving Query memory_search — memory_search("") for semantic search Ambiguous results — Surface all candidates to user, ask which one No results — Tell user nothing found, offer to create entry ### Disambiguation Protocol When user references something ambiguous (e.g., "John"): Search brain for all matches using memory_search("John") If multiple results: list them with context Found 2 matches for "John": 1. John Smith (Symph colleague, met 2024) 2. John Doe (GeeksOnABeach speaker, met 2026) Which one? Wait for confirmation before updating ### Templates Templates live in skills/brain/templates/. Each has: YAML frontmatter with structured fields Markdown body with standard sections When creating a new entry: Read the appropriate template Fill in known fields Leave unknown fields empty or with placeholder Write to brain//.md ### Categories Reference CategoryFolderUse ForPeoplebrain/people/Anyone user has met or wants to rememberPlacesbrain/places/Restaurants, landmarks, venues, locationsGamesbrain/games/Video games — status, opinions, notesTechbrain/tech/Devices, products, specs, quirksEventsbrain/events/Conferences, meetups, gatheringsMediabrain/media/Books, shows, films, podcastsIdeasbrain/ideas/Business ideas, concepts, random thoughtsOrgsbrain/orgs/Companies, communities, groups ### Linking Entities Use wikilink-style references to connect entities: [[people/raven-duran]] — link to a person [[events/geeksonabeach-2026]] — link to an event [[orgs/symph]] — link to an org This makes relationships explicit and searchable. ### Example Workflow User says: "Hey, I just met this guy called Raven Duran. He's positioning himself as an Agentic coder, met him at GeeksOnABeach PH last February." Agent does: memory_search("Raven Duran") → no results Read skills/brain/templates/person.md Create brain/people/raven-duran.md with filled template Optionally check/create brain/events/geeksonabeach-ph-2026.md and link User says: "The Raven Duran guy, he's still 26 years old" Agent does: memory_search("Raven Duran") → finds brain/people/raven-duran.md Read file via memory_get("brain/people/raven-duran.md"), update age: 26 in frontmatter Add note: - **2026-02-21**: Confirmed still 26 years old Update last_updated field ### Attachments Brain entries can have attachments: photos, PDFs, videos, audio, transcripts, etc. ### 🚨 MANDATORY: Save All Media Files When user sends ANY media (photos, audio, video, PDF) related to a brain entry: ALWAYS save the actual file to attachments/ — this is NON-NEGOTIABLE THEN analyze/transcribe the content into the profile NEVER skip saving the file just because you processed its content "Saved" means the FILE exists in attachments/, not just that content was transcribed. # REQUIRED: Copy the file cp /path/to/inbound/media.jpg brain/places/entry/attachments/descriptive-name.jpg If you transcribed content but didn't save the file → YOU DID IT WRONG. Go back and save it. ### Structure Flat file (no attachments): brain/places/manam.md Folder structure (with attachments): brain/places/mamou-prime-sm-podium/ mamou-prime-sm-podium.md # Profile (keeps original name) attachments/ index.md # Describes each attachment menu-page-1.jpg menu-page-2.jpg receipt.pdf storefront.mp4 ### Attachments Index (attachments/index.md) # Attachments | File | Description | Added | |------|-------------|-------| | menu-page-1.jpg | Menu first page, mains section | 2026-02-21 | | menu-page-2.jpg | Menu second page, desserts | 2026-02-21 | | receipt.pdf | Receipt from Feb visit, ₱2,400 | 2026-02-21 | | storefront.mp4 | Quick video of the entrance | 2026-02-21 | QMD (if enabled) indexes this file, making attachments searchable by description. ### Adding Attachments When user sends media about an entity (e.g., "Here's the menu for Mamou Prime"): Find the entry — memory_search("Mamou Prime") → brain/places/mamou-prime-sm-podium.md Convert to folder structure (if flat file): # Create folder mkdir -p brain/places/mamou-prime-sm-podium/attachments # Move profile into folder mv brain/places/mamou-prime-sm-podium.md brain/places/mamou-prime-sm-podium/ # Create attachments index touch brain/places/mamou-prime-sm-podium/attachments/index.md Save media to attachments/ with descriptive filename Update attachments/index.md with file description ### ⚠️ Always Save Original Files Do BOTH: Analyze/transcribe the content → add processed text to the profile (e.g., menu tables, business card info, transcript) Save the original files → preserve in attachments/ The text is searchable and processable. The originals are preserved artifacts. Never discard attachments unless user explicitly says "cleanup", "remove", or "delete" the files. Example: User sends menu photos ✅ Transcribe menu into markdown tables in profile ✅ Save original photos to attachments/menu-1.jpg, menu-2.jpg ✅ Update attachments/index.md Wrong: Only transcribing without saving originals ### Naming Attachments Be descriptive — the index provides context: menu-1.jpg, menu-2.jpg business-card.jpg product-demo.mp4 meeting-transcript.md voice-memo-2026-02-21.mp3 ### Example: Adding Menu Photos User sends: 2 photos with message "Menu at Mamou Prime" Agent does: Find brain/places/mamou-prime-sm-podium.md via memory_search("Mamou Prime") Convert to folder structure (if needed) Analyze photos → transcribe menu items, prices into markdown tables Update profile with transcribed menu section Save original photos as attachments/menu-1.jpg, attachments/menu-2.jpg Update attachments/index.md: # Attachments | File | Description | Added | |------|-------------|-------| | menu-1.jpg | Menu page 1 (transcribed to profile) | 2026-02-21 | | menu-2.jpg | Menu page 2 (transcribed to profile) | 2026-02-21 | Confirm to user: "Transcribed menu and saved 2 photos to Mamou Prime" ## Trust - Source: tencent - Verification: Indexed source record - Publisher: coderaven - Version: 1.3.0 ## Source health - Status: healthy - Item download looks usable. - Yavira can redirect you to the upstream package for this item. - Health scope: item - Reason: direct_download_ok - Checked at: 2026-04-22T14:04:09.960Z - Expires at: 2026-04-29T14:04:09.960Z - Recommended action: Download for OpenClaw ## Links - [Detail page](https://openagent3.xyz/skills/2nd-brain) - [Send to Agent page](https://openagent3.xyz/skills/2nd-brain/agent) - [JSON manifest](https://openagent3.xyz/skills/2nd-brain/agent.json) - [Markdown brief](https://openagent3.xyz/skills/2nd-brain/agent.md) - [Download page](https://openagent3.xyz/downloads/2nd-brain)