caveman
claudeTalk like caveman. Cut filler. Keep technical details explicit.
22 skills · 1 command · 3 agents · hooks
Persistent, git-versioned, human-gated memory for coding agents.
Plugin hipocampo from marketplace hipocampo in Ferraz2000/hipocampo-memory. Commands read from the repo's manifest — 1 month ago.
archive-closedCompact long-closed insights into a one-line archive index so they stop costing discovery context while staying traceable. Use when the user says "/archive-closed", "compact the insights", "clean up old insights".
auditFact-check a vault page against its cited sources — every claim must trace to a source that actually supports it. Use when the user says "/audit <path>", "fact-check this note", "verify this page against its sources".
brain-initScaffold a hipocampo memory vault into the current project — creates brain.config.toml and the docs/brain skeleton (knowledge/index, capture.md, context-budget.md, note templates). Use when setting up agent memory in a repo that doesn't have a vault yet, or when the user says "init the brain", "set
brain-router-initGenerate a lean, operational AGENTS.md router for this repo (any language) plus the CLAUDE.md pointer and a .claude/rules/USER.md persona seed. Use after brain-init, or when the user says "generate the router", "set up AGENTS.md", "brain-router-init".
brain-scripts-initVendor the hipocampo validation scripts, git hooks, and CI workflow into the current repo so the doc-sync gate and preflight run locally and in CI. Use after brain-init, or when the user says "vendor the scripts", "set up the hooks/CI", "brain-scripts-init".
brain-updateUpdate the vendored hipocampo scripts/hooks/CI in this repo to the kit's current version without clobbering local changes or any vault content. Use when the user says "update hipocampo", "bump the kit", "brain-update".
captureCapture the current conversation into durable, git-versioned memory (source / concept / contract / proposal / persona) with a human write-gate. Use when the user says "/capture", "capture this", "save to the brain", or pt-BR phrases like "registra isso" / "salva no brain", "remember this decision",
challengeBefore committing to a decision or approach, confront it with the vault's past reversals, failures, and superseded decisions on the same topic. Use when the user says "/challenge", "has this been tried", "did we decide against this", or before a significant architectural choice.
discover-standardsMine the existing codebase for de-facto conventions (naming, structure, error handling, test patterns) and propose them as candidate insights for the vault. Use when the user says "/discover-standards", "extract our conventions", "what patterns do we follow".
discoveryDiscovery mode — authorize broad, bounded reading of the vault for audits, cross-cutting insights, or roadmap review. The declared exception to never-bulk-read. Use when the user says "/discovery", "explore the vault", "do discovery on area X", "audit Y".
execute-insightExecute exactly one approved insight from the vault, with separate commits for the code and for closing the insight's frontmatter. Use when the user says "/execute-insight <path-or-slug>", "execute this insight", "implement insight X".
from-roadmapPull the next actionable insight from the vault work queue (active/triage, unblocked, best impact-vs-effort) and propose its execution. Use when the user says "/from-roadmap", "what should I work on", "next item".
gardenRun a gardener pass over the vault — find contradictions, stale claims, orphan sources, and missing cross-links, then propose fixes. Use when the user says "/garden", "tidy the vault", "lint the knowledge base", or periodically.
implementImplement an already-approved spec from the vault (specs/), staying inside its declared scope and acceptance criteria. Use when the user says "/implement <spec>", "implement the spec", "build what the spec says".
low-tokenLean execution mode for small, well-scoped tasks — minimum reading, no vault, no discovery, smallest correct change. Use when the user says "/low-token <task>", "lean mode", "minimal context".
postmortemMine the just-finished branch/PR for durable lessons and propose them as vault captures (insights or knowledge) — the user picks, the agent files. Use when the user says "/postmortem", "capture lessons from this branch", "what did we learn".
promotePromote an insight into a durable artifact — a spec, an ADR, or a tracker issue — and mark the insight promoted. Use when the user says "/promote <insight> <spec|adr|issue>", "turn this insight into a spec/ADR/issue".
recallSearch durable memory (the vault) mid-task and pull only the relevant notes into context, instead of asking the user about past decisions. Use proactively before asking "what did we decide about X", when you need prior rationale, conventions, or history; or when the user says "/recall", "recall", "w
reflectIterate on a work product (code, plan, spec, summary) with an in-session generate→critique→revise loop bounded by config stopping criteria, seeded with past lessons from memory and closed by capturing the distilled lesson. Use when the user says "/reflect", "iterate on this", "critique and improve",
searchKeyword-search the vault (knowledge/insights/specs/raw sources) by relevance, BM25 with an optional FTS5 fast path. Use when the user says "/search", "search the brain/vault", "find notes about X".
specCreate a short, approved-style spec (mini design doc) in the vault from the template — no code. Use when the user says "/spec", "write a spec for X", "design doc for Y".
weeklyWeekly review cycle over the vault — triage new insights and inbox sweeps, surface stale/expired pages, and propose promotions or closures. Use when the user says "/weekly", "weekly review", "triage the vault".
Talk like caveman. Cut filler. Keep technical details explicit.
22 skills · 1 command · 3 agents · hooks
Persistent memory for Claude Code: the next session starts where the last one left off.
29 skills · 1 MCP server · hooks
tasteskill 2.5.1: professional company websites at the figma.com and x.ai level, with one section per screen, real photographs, real product UI, effects kept for tech and gaming pages, and a real-browser checker and finish gate.
1 skill
AI-powered codebase understanding — analyze, visualize, and explain any project
9 skills · 10 agents · hooks
Open-source AI job search that runs locally in your AI coding CLI: scan job portals, evaluate listings into a structured A-H report with a 1-5 score, tailor your CV, track applications
5 skills · hooks
Cross-session memory and token savings for coding agents.
6 skills · 1 agent · 1 MCP server · hooks
This page updates itself from your repo. Add the badge to your README, or get it Featured to pin it above the organic results with a followed link to your site.