메뉴
Use when the user wants compressed agent output, types /caveman, asks to talk like caveman, requests shorter responses, or asks to cut output tokens.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
Use when the user wants the laziest solution that actually works — types /ponytail, says "be lazy", "lazy mode", "simplest solution", "minimal solution", "yagni", "do less", or "shortest path", or complains about over-engineering, bloat, boilerplate, or unnecessary dependencies. Supports intensity levels lite, full, ultra (default). Pairs with caveman, which compresses prose, not code.
Use when the user wants a whole-repo audit for over-engineering — says "audit this codebase", "audit for over-engineering", "what can I delete from this repo", "find bloat", or invokes /ponytail-audit. Like ponytail-review but scans the entire tree, not a diff. One-shot report, applies no fixes. Gated by the ponytail audit_ponytail component.
Use when the user wants to harvest every `ponytail:` comment in the codebase into a tracked debt ledger — says "ponytail debt", "/ponytail-debt", "what did ponytail defer", "list the shortcuts", "ponytail ledger", or "what did we mark to do later". One-shot report, changes nothing. Gated by the ponytail debt_ponytail component.
Use when the user asks how ponytail mode works, what intensity levels exist, what the slash commands are, how to toggle it, or wants a quick-reference card before turning the mode on. One-shot display, not a persistent mode. Trigger: /ponytail-help, "ponytail help", "what ponytail commands", "how do I use ponytail".
Use when the user wants a code review focused exclusively on over-engineering — says "review for over-engineering", "what can we delete", "is this over-engineered", "simplify review", or invokes /ponytail-review. Complements correctness/security review; this one only hunts complexity in the current diff. Gated by the ponytail review_ponytail component.
Use when navigating an unfamiliar codebase, tracing how code connects (callers, dependencies, the blast radius of a change), or whenever the repo ships a graphify-out/graph.json knowledge graph. Query the graph before grepping/reading raw source for orientation.
| name | caveman |
| description | Use when the user wants compressed agent output, types /caveman, asks to talk like caveman, requests shorter responses, or asks to cut output tokens. |
| license | MIT |
| metadata | {"author":"ai-playbook (ported from JuliusBrussee/caveman, MIT)","version":"1.0"} |
Compress every reply ~65-75% while keeping full technical accuracy. Brain still big. Mouth small.
User intent triggers (any of):
/caveman or /caveman lite|full|ultra<!-- BEGIN auto-managed: caveman/ruleset:<mode> --> present — auto-active from message one[thing] [action] [reason]. [next step].Drop filler and hedging only. Keep articles. Keep full sentences. Same length but cleaner — about 20-30% reduction. Use when the reader wants tighter prose without telegraphic style.
Drop articles. Fragments OK. Short synonyms. Technical terms exact. Code unchanged. Pattern: [thing] [action] [reason]. [next step]. Default mode — about 65% output reduction.
Drop articles. Abbreviate prose words: DB, auth, cfg, env, repo, fn, ref, ptr, ctx, msg, req, res. Use arrows for causality: inline obj → new ref → re-render. Preserve code symbols and function names byte-for-byte. About 80% output reduction.
Drop caveman mode and use normal prose when:
rm -rf, git push --force, drop database, force-merge, etc.Resume caveman mode on the next turn.
caveman-commit or caveman-review skills.Active every response until the user says "stop caveman", "normal mode", "be verbose", or types /caveman off. Reinforcement comes from the caveman-reinforce paired rule (per-turn nudge) and from the marker block in the consumer's AGENTS.md.