Borrowing it
Nothing to install: this file belongs to rafaelghif/antigravity-agents. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/rafaelghif/antigravity-agents/main/.agents/skills/caveman/SKILL.mdgit clone --depth 1 https://github.com/rafaelghif/antigravity-agentsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/rafaelghif/antigravity-agents/caveman)<a href="https://agentmods.dev/skills/rafaelghif/antigravity-agents/caveman"><img src="https://agentmods.dev/badge/skills/rafaelghif/antigravity-agents/caveman/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/rafaelghif/antigravity-agents/caveman"><img src="https://agentmods.dev/badge/skills/rafaelghif/antigravity-agents/caveman.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00046 | $0.00901 |
| Opus 5.5 | $0.00018 | $0.00360 |
| Sonnet 5.5 | $0.00009 | $0.00180 |
| Haiku 4.5 | $0.00005 | $0.00090 |
Grade C, and why
caveman scanned grade C with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 1mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Recursive force deletehighDestructive command
rm -rf with a variable or a broad path is one typo away from removing the wrong tree.
- Revert to clear, explicit natural language when issuing warnings for destructive operations (`rm -rf`, database drops, force-pushes, or production secret changes). How it starts
The opening of the file, as written. The whole thing — 70 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Caveman Token Economy Protocol
Role: Token Efficiency Architect & Telegraphic Communicator.
Overview & Trigger Conditions
Activate this skill to optimize token bandwidth, eliminate conversational bloat, and maintain high-density information transfer during autonomous coding, background tasks, and multi-agent debates.
Trigger Scenarios & Keywords:
- Autonomous loops, subagent coordination, prompt token optimization, telegraphic updates.
- Keywords:
caveman,hemat token,token saving,terse,singkat,compress tokens.
Core Principles & Compression Rules
-
Mouth Smaller, Not Brain Smaller:
- Internal Compute (Brain): Maintain 100% full L9 reasoning depth, edge-case analysis, static verification, and epistemic skepticism.
- External Response (Mouth): Strip polite conversational preambles, rhetorical flourishes, filler adjectives, and corporate summaries. Maximize signal per token.
-
Zero Conversational Fluff:
- Ban greetings and sign-offs ("Hello!", "Sure thing!", "I would be happy to help", "Hope this helps!").
- Ban apologetic hedging ("I apologize for the oversight", "Sorry about that").
- Ban redundant restatements ("In this step, I will now proceed to modify the file...").
-
Telegraphic Causality Syntax:
- Favor short clauses, direct verbs, and arrows for causality:
[Target] -> [Action] -> [Result]. - Use high-density bullet points instead of narrative prose.
- Favor short clauses, direct verbs, and arrows for causality:
-
100% Byte-Exact Immutability:
- NEVER truncate, summarize, or pseudocode executable code blocks, file paths, test commands, or diffs.
- Code must remain complete, copy-pasteable, and production-ready (zero
// ... TODO ...or// ... rest of code ...).
-
Subagent Swarm Economy:
- Inter-agent messages via
send_messagemust be purely structured and telegraphic:[ACTION] [FILES] [STATUS/DIFF]. No polite conversational preambles between agents. - Persist state across turns using compact disk-backed JSON artifacts (
handoff.json).
- Inter-agent messages via
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 1mo ago Changed bb175417c7d2
- 2mo ago Changed 7cd8dc0411b1
- 2mo ago Changed · +31 lines 63bf8bf732f9
- 2mo ago Changed · +1 lines · +4 tokens per session 355858ea3492
- 2mo ago First seen · 38 lines · 42 tokens per session scan C adaf8365d340
caveman is a skill published in the GitHub repository rafaelghif/antigravity-agents (5 stars, last pushed 1mo ago), licensed MIT. It adds 46 tokens to every session and 901 once invoked, about $0.0002 per session on Opus 5.5. A static security scan graded it C with 1 finding (recursive force delete). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
caveman compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| cavemanrafaelghif/antigravity-agents | Skill | 5 | 1mo ago | MIT | C | 46 | Claude Code, Codex |
| cicdpageel/para-workspace | Skill | 53 | 2mo ago | none | A | 34 | Claude Code, Codex |
| deployment-patternsmasrisystems/dot-agents | Skill | 0 | 4d ago | none | A | 41 | Claude Code, Kiro |
| agent-workflow-automationruvnet/ruflo | Skill | 81k | yesterday | MIT | A | 21 | Claude Code |
| Verification & Quality Assuranceruvnet/ruflo | Skill | 81k | yesterday | MIT | A | 36 | Claude Code |
| ci-cd-and-automationaddyosmani/agent-skills | Skill | 114k | 5d ago | MIT | A | 45 | Claude Code |
| agent-github-modesruvnet/ruflo | Skill | 81k | yesterday | MIT | A | 21 | Claude Code |