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 tmj-90/gaffer --skill card-reviewgit clone --depth 1 https://github.com/tmj-90/gafferWrote 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/tmj-90/gaffer/card-review)<a href="https://agentmods.dev/skills/tmj-90/gaffer/card-review"><img src="https://agentmods.dev/badge/skills/tmj-90/gaffer/card-review/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/tmj-90/gaffer/card-review"><img src="https://agentmods.dev/badge/skills/tmj-90/gaffer/card-review.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 | $0.00125 | $0.00846 |
| Opus 5 | $0.00063 | $0.00423 |
| Sonnet 5 | $0.00025 | $0.00169 |
| Haiku 4.5 | $0.00013 | $0.00085 |
Grade A, and why
card-review 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 9d 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.
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 — 86 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Review one file card — the semantic accuracy gate
You will be given:
- A card (tldr + role) produced by the generation pass.
- The mechanical structure of the file (imports, top-level symbol names).
- A bounded head snippet of the file.
Your job: judge whether the TLDR and role are directionally accurate and not over-claiming — given ONLY the structure and snippet shown.
You are NOT asked to rewrite the card. You are asked to give a verdict and one sentence explaining it.
The accuracy bar
This is a generous but honest bar. You are not looking for perfection; you are catching meaningful errors.
A card PASSES if:
- The TLDR's primary claim is consistent with the structure and snippet.
- Imprecision or vagueness is acceptable as long as nothing is actively wrong.
- The role label is plausible (even if another label might be slightly better).
A card REVISES if:
- The TLDR is partially right but contains a specific false claim that could
mislead an agent (e.g. claims it exports
handleSessionwhen no such symbol appears in the structure). - The role label is clearly wrong (e.g.
migrationfor a route handler).
A card REJECTS if:
- The TLDR describes something fundamentally different from what the file is (e.g. "implements JWT authentication" for a file that only does arithmetic).
- The TLDR is so vague it provides zero retrieval value ("utility functions").
- The TLDR contains fabricated behaviour not visible in the structure or snippet.
What you must NOT do
- Do NOT penalise a card for being brief or imprecise — only for being wrong.
- Do NOT infer capabilities from the filename alone; use only what the structure and snippet show.
- Do NOT reject a card because you would have written it differently. The bar is directional accuracy, not editorial quality.
- Do NOT consider the symbols field in your verdict — that is the deterministic gate's job. Focus only on the TLDR and role_primary.
- Do NOT write a revised TLDR. Your output is verdict + one-sentence reason only.
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.
- 9d ago First seen · 86 lines · 125 tokens per session scan A 89092516398c
card-review is a skill published in the GitHub repository tmj-90/gaffer (2 stars, last pushed 2d ago), licensed Apache-2.0. It adds 125 tokens to every session and 846 once invoked, about $0.0006 per session on Opus 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.
Other skills, from other repositories
code-review
Comprehensive code review with distinct aspect based sections. Use when reviewing code, checking for security issues, finding type safety problems, auditing code quality, or when user asks to review code, PRs or changes. Three-phase workflow runs static tools, LLM judgment, and writes diagnostic log.
test-audit
Audit test suites for T1-T4 violations using AST analysis, mock detection, and multi-stage synthesis. Invoke when user asks to audit tests, check test quality, find mock violations, review test effectiveness, or inspect test suites for over-mocking. Triggers automatic rewrites when quality gates fail.
spec-drift-check
Audits a WP brief for drift against current code — extracts claims, verifies each, emits PROCEED/STOP verdict. Use when starting a new WP, before consuming a spec as binding, or when a doc references paths/lines/functions.
besimple-broccoli-blind
Non-interactive wrapper: plan-sketch -> auto-pick recommended options -> plan-write -> plan-critique-loop -> implement-from-plan -> claude-simplify-wrapper -> dedup -> code-review-loop. No PR creation and no Linear comments.
code-review-loop
Iterative review+fix loop for BASESHA..HEAD: generate findings, apply accepted fixes, run checks, commit, and re-review up to 3 iterations or until clean.
dedup
Dedupe-only pass for BASESHA..HEAD: remove duplicate code introduced by the diff or reuse existing shared utils; applies changes + commits.