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 agentmods add commands/solanabr/auditor-skill/re-auditgit clone --depth 1 https://github.com/solanabr/auditor-skillWrote 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/commands/solanabr/auditor-skill/re-audit)<a href="https://agentmods.dev/commands/solanabr/auditor-skill/re-audit"><img src="https://agentmods.dev/badge/commands/solanabr/auditor-skill/re-audit.svg" alt="Measured on agentmods" 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 | $0.00114 | $0.01658 |
| Opus 5 | $0.00057 | $0.00829 |
| Sonnet 5 | $0.00023 | $0.00332 |
| Haiku 4.5 | $0.00011 | $0.00166 |
Grade A, and why
auditor:re-audit 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 5d 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 — 39 lines — stays where its author put it; the contents beside it link to each section on GitHub.
auditor-skill — Fix Review / Re-Audit
Arguments: $ARGUMENTS
Re-audit a codebase after a fix round. Input: a prior audit_<n>/REPORT.md (arg 1) and the current tree; the delta range is base..head (arg 2, default main..HEAD — the fix commits). Read OUTPUT-RULES.md first (severity 1-10, the Rule 5b gate). This is a Zenith/Certora-style fix-review, not a fresh audit — it reuses the prior report as the baseline and reports what changed.
Steps
-
Load the baseline. Parse the prior report's findings (id, severity,
file:line, root-cause, Rule 5b block). Pin both commits: the prior report's audited commit and the currenthead. -
Classify every prior finding. For each prior finding, read the current code at the (possibly moved) location and assign one of six states. Every non-open status must cite the remediation commit/PR hash that changed it (
git log --oneline/ the PR the fix landed in), not merely the currentfile:line— the reader must be able to point at what fixed it, not just where the code now sits.- FIXED — the vulnerable path is gone. Re-run Rule 5b on the fix: confirm the fix actually closes the reachability/bound the finding claimed — a cosmetic or incomplete patch is not FIXED. Cite the fixing commit/PR hash and
file:line. - PARTIALLY-FIXED — the fix closes some of the reachability/bound but a residual path survives (one call site patched, an edge input still breaks it, the guard added is necessary-but-insufficient). Re-run Rule 5b on the residual path and carry a severity for what remains. Cite both the fixing commit/PR hash and the surviving
file:line. - STILL-OPEN — the finding survives unchanged (or was never touched). Carry its severity forward.
- REGRESSED — the fix introduced a new problem, or a previously-fixed issue returned, or the "fix" made it worse. Treat as a fresh finding through the full Rule 5b gate; a security check deleted in a fix commit is a CRITICAL regression (git-blame it, cite the commit).
- ACKNOWLEDGED — the client accepts the risk and has not changed the code (won't-fix / risk-accepted). The code still contains the vulnerable path, so severity is unchanged — do not score it as resolved; record it as an accepted-risk item.
- DISPUTED — the client contests the finding's validity/severity and the code is unchanged. Record the dispute as the client's position, then state our independent re-derivation verdict (upheld / downgraded / withdrawn) from the code — the report keeps our verdict, not the client's assertion, as the ruling.
- FIXED — the vulnerable path is gone. Re-run Rule 5b on the fix: confirm the fix actually closes the reachability/bound the finding claimed — a cosmetic or incomplete patch is not FIXED. Cite the fixing commit/PR hash and
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.
- 5d ago First seen · 39 lines · 114 tokens per session scan A 1d5827337840
auditor:re-audit is a command published in the GitHub repository solanabr/auditor-skill (52 stars, last pushed 1mo ago), licensed MIT. It adds 114 tokens to every session and 1,658 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 2026-08-30.
Other commands, from other repositories
wp-perf-review
WordPress performance code review - detects database anti-patterns, caching issues, hook problems, and scalability concerns.
wp-perf
Quick WordPress performance scan - fast triage using critical pattern detection.
genshijin-compress
Markdown・テキストファイルを原始人形式へ安全圧縮.
save
Save this conversation as a new or existing reusable context.
valuation-methods
Valuation methods analysis — multiples, DCF inputs, PEG integration, valuation assumption extraction.
enum-udp
UDP scan + service follow-up — top ports first, full sweep only when justified.