The Monkey Brain

Plugin brain from marketplace monkey-brain in NeamulMorshed/The-Monkey-Brain. Commands read from the repo's manifest — 4 days ago.

Skills (24)

• brief

  Pull a cited context pack (at most ~2k tokens) on a topic from the project's Monkey Brain. Use when the user says "brief me on X" or "catch me up on X", or before work that past decisions, specs or research may already cover.

• career

  Private career pack in .brain/private/ (never committed or indexed) — case studies from shipped work, CV variants, an evidence-backed skill matrix, mock interviews. Use for "case study", "mock interview", "update/tailor my CV" or "skill matrix".

• ci

  Add a GitHub Actions CI workflow from the detected stack (Node, Python, Go, .NET, Rust) running the project's own lint, test and build steps; never overwrites without asking. Use for "set up CI", "add a CI pipeline", or when doctor reports no CI.

• compress

  Permanently rewrite an instruction or memory file (CLAUDE.md, memory notes, rules pages) into a terser form with identical meaning, cutting its per-session token cost. Use for "compress <file>" or a bloated CLAUDE.md. Not for wiki pages whose wording is evidence.

• dashboard

  One-page offline HTML dashboard of this project's Monkey Brain — index stats, open specs with AC progress, loops, health, recent log and decisions, token usage — written to .brain/sessions/. Use for "dashboard" or "show me the brain".

• digest

  Standup or weekly review from the Monkey Brain — what's blocked, what got done (log and commits), what's in flight — filed to sessions/. Use for "standup", "daily brief", "what did we do yesterday" or "weekly review".

• doctor

  Run the 21-check health report on the project's Monkey Brain (links, orphans, index, log gaps, hooks, budget, open P0s, schema, cache, CI) and triage the fixes. Use for "brain doctor", "check brain health", "is the brain healthy", or before a release.

• dump

  File a loose note where it belongs in the Monkey Brain — a decision as an ADR, a fact in memory/, a next step on its workstream, an idea, a link in Clippings/, a repeated correction as an instinct. Use for "dump", "jot this down", "we decided …".

• game

  Game-development pipeline — concept, GDD, prototype spec, build, playtest, balance ADRs — reusing plan, build and wrap. Use when the user starts a game, writes a GDD, designs mechanics or a core loop, logs a playtest, or tunes balance.

• home

  One dashboard across every Monkey Brain project on this machine — health, open specs, loops, token usage per project and combined. Use for "show all my brains", "dashboard for all projects" or "monkey brain home". Needs no .brain/ here.

• ingest

  Compile a source into the project's Monkey Brain wiki (the 8-step ingest). Use when the user says "ingest this", shares an article, doc, URL or paste to add to the brain, or asks to process Clippings.

• init

  Scaffold a Monkey Brain (.brain/) into the current project, or refresh an existing brain's schema without touching its knowledge. Use when the user asks to set up, initialize or create a brain, or accepts the session-start offer.

• learn

  Learning pack — spaced-repetition drills (SM-2) plus one new concept per session, compiled into the brain; only due cards enter context. Use for "practice <topic>", "drill my cards", "flashcards" or "spaced repetition".

• lint

  Health-check the Monkey Brain wiki — a mechanical scan (broken links, orphans, frontmatter, index drift, strays), then reasoning over the flagged pages, fixes applied. Use for "lint the brain", "lint the wiki", or after a large ingest.

• lock

  Take, check or release the team work lock (a committed, expiring LOCK.md) on the whole brain or one spec, so two people don't change the same thing. Use for "lock the brain", "lock the <x> spec", "release the lock" or "who has the lock".

• loop

  Run a bounded loop that stops on the brain's own criteria — build until every AC passes, research until the recommendation is stable, design until no P0 — with tick caps and stall and livelock halts. Use for "loop until done" or "keep going until green".

• plan

  Turn a feature idea into specs/<feature>.md with numbered acceptance criteria and a tier — the record the plan and TDD gates read. Use for "spec X", "draft a spec", "plan the feature", or before non-trivial work. Only the curator approves a plan.

• product-design

  Product-design process end to end — discovery, definition, ideation, design, validation — with every artifact filed into the brain. Use for designing a product, feature or UX, personas, journeys, JTBD, how-might-we, usability or accessibility checks.

• query

  Answer a question from the project's Monkey Brain — index-first retrieval with citations, novel answers filed back as synthesis pages. Use when the user asks what the brain knows, wants an overview or comparison, or says "ask the brain".

• research

  Investigate a topic and file the findings into the Monkey Brain — wiki, then codebase, then web — as a wiki/research/ page with sources and a recommendation. Use for "research X", comparing options before building, or when a plan needs evidence.

• review

  Verify a built spec AC by AC with evidence, review its code, file the findings back (review page, ADRs, instincts) and close the spec honestly. Use for "review the changes/spec/branch/PR", after /brain:build, or before merging feature work.

• terse

  Terse output mode (Caveman-style, ~65% shorter), on by default; this skill turns it off or on. Code, commands, paths and errors never compress. Use for "terse off", "be more verbose", "be terse", "shorter answers", or verbosity complaints.

• usage

  Real token usage for this project from Claude Code's transcripts — per day, model and branch, subagent share, cache-hit ratio — and what to change. Use for a token report, token spend, the cache-hit ratio, or why sessions feel expensive.

• wrap

  Definition-of-done for a work session — verify, sync the bookkeeping (log, index, resume), commit with the vault's conventions. Use for "wrap up", "end the session", "call it a day", or after a milestone.

Subagents (2)

• brain-librarian

  Routine Monkey Brain maintainer for batch and bookkeeping work. Delegate ingesting ONE source (run several in parallel for a batch) — it performs the ingest compile (summary page, cross-links, index + log updates) in an isolated window so the main session sees only the result. Used by /brain:ingest

• brain-researcher

  Read-only research worker for a Monkey Brain. Delegate ONE focused slice of a research question — a set of wiki pages, a codebase area, or a web topic — and it returns cited findings without writing anything. Spawn several in parallel (one per slice) for wide coverage at low context cost; the lead s

MCP servers (1)

• brain-search

  runs: node ${CLAUDE_PLUGIN_ROOT}/hooks/scripts/search-mcp.js

Hooks (15)

• SessionStart

  node "${CLAUDE_PLUGIN_ROOT}/hooks/scripts/brain-status.js"

• SessionStart

  node "${CLAUDE_PLUGIN_ROOT}/hooks/scripts/resume.js"

• UserPromptSubmit

  node "${CLAUDE_PLUGIN_ROOT}/hooks/scripts/trigger-router.js"

• UserPromptSubmit

  node "${CLAUDE_PLUGIN_ROOT}/hooks/scripts/recall.js"

• PreToolUse

  node "${CLAUDE_PLUGIN_ROOT}/hooks/scripts/guards.js"

• PreToolUse

  node "${CLAUDE_PLUGIN_ROOT}/hooks/scripts/agent-track.js"

• PostToolUse

  node "${CLAUDE_PLUGIN_ROOT}/hooks/scripts/wiki-check.js"

• PostToolUse

  node "${CLAUDE_PLUGIN_ROOT}/hooks/scripts/instinct-track.js"

• TaskCreated

  node "${CLAUDE_PLUGIN_ROOT}/hooks/scripts/resume-log.js"

• TaskCompleted

  node "${CLAUDE_PLUGIN_ROOT}/hooks/scripts/resume-log.js"

• Stop

  node "${CLAUDE_PLUGIN_ROOT}/hooks/scripts/wrap.js"

• SubagentStop

  node "${CLAUDE_PLUGIN_ROOT}/hooks/scripts/agent-track.js"

• PreCompact

  node "${CLAUDE_PLUGIN_ROOT}/hooks/scripts/snapshot.js"

• SessionEnd

  node "${CLAUDE_PLUGIN_ROOT}/hooks/scripts/resume-log.js"

• SessionEnd

  node "${CLAUDE_PLUGIN_ROOT}/hooks/scripts/wrap.js"

Similar plugins