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 akasecurity/claude-tools --skill shell-auditgit clone --depth 1 https://github.com/akasecurity/claude-toolsWrote 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/akasecurity/claude-tools/shell-audit)<a href="https://agentmods.dev/skills/akasecurity/claude-tools/shell-audit"><img src="https://agentmods.dev/badge/skills/akasecurity/claude-tools/shell-audit/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/akasecurity/claude-tools/shell-audit"><img src="https://agentmods.dev/badge/skills/akasecurity/claude-tools/shell-audit.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 | $0.00112 | $0.00606 |
| Opus 5.5 | $0.00045 | $0.00242 |
| Sonnet 5.5 | $0.00022 | $0.00121 |
| Haiku 4.5 | $0.00011 | $0.00061 |
Grade A, and why
shell-audit scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
What it actually says
shell-audit
Run the bundled read-only auditor and present its findings clearly and honestly.
Run it
bash "$CLAUDE_CONFIG_DIR/skills/shell-audit/audit.sh" [rc-file]
No argument → it picks the rc from $SHELL. It walks that rc and every file
reachable through its source / . chain (cycle-safe via a visited-set) and
prints four sections: credentials, persistence patterns, alias
hygiene, git drift.
Present the results
- Credentials first (highest severity). NEVER print secret values — the script redacts them; keep them redacted. Recommend moving any secret out of startup files into a keychain / credential helper, and rotating anything that was exposed (these files are often git-tracked and world-readable to the user).
- Persistence patterns = "eyeball these". Most are legit (setup scripts); for each, confirm it fetches/execs only what's intended and that nothing writes into another rc.
- Alias hygiene — offer to dedupe (decide which definition wins) and note any launcher alias whose target dir is missing.
- Git drift — confirm each change is the user's. A modified-vs-HEAD or untracked startup file is the strongest tamper signal available here.
- Partial coverage — if the run prints "COVERAGE IS PARTIAL", a
sourceline used a variable the auditor couldn't resolve without running the shell (e.g.source "$DOTFILES/x"); that file was not audited. Surface it and offer to re-run pointed at the resolved path. - State the scope limit every time: this covers the shell-startup slice only — not launchd/LaunchAgents, cron, login items, ssh authorized_keys/config, or git hooks. Don't present it as comprehensive persistence detection.
Don't
- Don't modify any startup file as part of the audit — it's read-only. If the
user wants fixes, propose them and let the user apply: their dotfiles are
Edit/Write-denied by design, so hand edits over via a
! <cmd>prompt.
What ships with it
1 file 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.
- 2mo ago First seen · 48 lines · 112 tokens per session scan A 231af26dc8b8
shell-audit is a skill published in the GitHub repository akasecurity/claude-tools (11 stars, last pushed 10d ago), licensed MIT. It adds 112 tokens to every session and 606 once invoked, about $0.0004 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.
Other skills, from other repositories
blueprompt
Create app blueprints and agent prompts from rough ideas. Use when the user says "blueprompt" or "app blueprompt". Transforms concepts into structured specifications for v0, Lovable, Replit, or any AI builder.
atlas
Architect the intelligence layer for agentic systems — RAG pipelines, model selection, embeddings, evaluation, and knowledge systems. Use when the user says "atlas", "ai data", "data arc". Produces data/ML architecture blueprints.
angel
Design and protect scalable backend systems with clean contracts, strong boundaries, and reliability. Use when the user says "angel", "angel backend", "backend strat" or "backend design". Produces architecture recommendations, refactor plans, and maintenance strategies.
argus
Design security architectures for agentic applications. Use when the user says "argus", "agent sec", or "agent auth". Creates layered defenses, audit systems, authentication patterns, and resilience safeguards.
bliss
Create distinctive, high-quality frontend interfaces with high design quality. Use when the user says "bliss", "bliss ui", or "bliss interface". Generates creative, accessible, polished code that avoids generic AI aesthetics.
clarity
Human-first UX strategy. Solve the right problem before solving it beautifully. Use when the user says "clarity", "clarity ux", "ux research". Produces evidence-based UX artifacts that prepare for visual design.