Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add HigorAlves/orc --skill caveman-prgit clone --depth 1 https://github.com/HigorAlves/orcWrote 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/higoralves/orc/caveman-pr)<a href="https://agentmods.dev/skills/higoralves/orc/caveman-pr"><img src="https://agentmods.dev/badge/skills/higoralves/orc/caveman-pr/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/higoralves/orc/caveman-pr"><img src="https://agentmods.dev/badge/skills/higoralves/orc/caveman-pr.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 | $0.00014 | $0.01230 |
| Opus 5.5 | $0.00006 | $0.00492 |
| Sonnet 5.5 | $0.00003 | $0.00246 |
| Haiku 4.5 | $0.00001 | $0.00123 |
Grade A, and why
caveman-pr scanned grade A with 0 findings 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 yesterday.
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.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 88 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Write PR descriptions terse and exact. Reviewers don't need a tour of your diff — they need a why, a how-tested, and (if relevant) a heads-up.
Rules
Title:
- Same shape as the squash-merge commit subject — Conventional Commits if the project uses them (
feat(scope): ...,fix(scope): ...). - ≤ 72 chars; ≤ 50 if possible.
- Imperative mood ("add", "fix", "remove" — not "added").
- No
[WIP]— use a draft PR instead. - No emoji unless the project already does it.
Body — at most 4 sections, in this order. Omit any that don't add signal.
Why (almost always mandatory)
- 1–3 sentences. Link the plan / RFC / ADR / Jira ticket / issue if that's the why.
- Skip ONLY when the title is genuinely self-explanatory (e.g.
fix(billing): guard against zero-cap users— title says everything).
What changed (often skippable)
- ONLY include if the diff isn't obvious from a 30-second skim.
- Bullets of structural changes. Don't restate file paths the reviewer will see anyway.
- Drop "Refactored", "Cleaned up" — say what behavior changed.
- When the shape is the point, one small visual beats bullets: a
diff-sketch of a call tree, component tree, or file tree, pseudocode of the new control flow, or a short mermaid sequence. One visual, never a gallery.
How tested (mandatory)
- One or two lines. Command run + result. Or a link to QA artifacts.
- Web change: link to
.orc/<branch>/files/qa/steps.md(or wherever orc's QA evidence lives). - "Tests: 47 pass, 0 fail." beats "Wrote tests for the new branch and they all pass."
- Evidence tiers: a screenshot or recording (S) when the change is visual and the environment allows it; execution output (A) — the exact test that failed before and passes now, command + result; prose on its own never counts.
Notes (only if there's something real)
- Merge danger — required for anything beyond a trivial fix, one line:
Door: one-way | two-way · Blast radius: <one or two words>. Two-way doors revert cleanly; one-way doors (migrations, deletions, published API changes, anything a revert won't undo) say what makes them hard to walk back. - Breaking changes — call out, including the migration step.
- Risks the reviewer must look at twice.
- Follow-up issues filed.
- Skip platitudes ("clean code", "all tests pass" — already covered).
What ships with it
2 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- yesterday Changed 7dc4a609e193
- 17d ago Changed · -52 lines · -34 tokens per session 16089c5f313e
- 2mo ago First seen · 140 lines · 48 tokens per session scan A 73b24ca1a8b4
caveman-pr is a skill published in the GitHub repository HigorAlves/orc (7 stars, last pushed yesterday), licensed MIT. It adds 14 tokens to every session and 1,230 once invoked, about $0.0001 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
caveman-pr compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| caveman-prHigorAlves/orc | Skill | 7 | yesterday | MIT | A | 14 | Claude Code |
| clean-code-cppCassetteTapeCrackle/claude-starters | Skill | 1 | 4d ago | MIT | A | 42 | Claude Code |
| clean-code-coreCassetteTapeCrackle/claude-starters | Skill | 1 | 4d ago | MIT | A | 70 | Claude Code |
| clean-code-rustCassetteTapeCrackle/claude-starters | Skill | 1 | 4d ago | MIT | A | 44 | Claude Code |
| clean-code-cCassetteTapeCrackle/claude-starters | Skill | 1 | 4d ago | MIT | A | 34 | Claude Code |
| clean-code-goCassetteTapeCrackle/claude-starters | Skill | 1 | 4d ago | MIT | A | 34 | Claude Code |
| clean-code-pythonCassetteTapeCrackle/claude-starters | Skill | 1 | 4d ago | MIT | A | 39 | Claude Code |