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 mirkobozzetto/arsenal --skill reviewgit clone --depth 1 https://github.com/mirkobozzetto/arsenalWrote 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/mirkobozzetto/arsenal/review)<a href="https://agentmods.dev/skills/mirkobozzetto/arsenal/review"><img src="https://agentmods.dev/badge/skills/mirkobozzetto/arsenal/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/mirkobozzetto/arsenal/review"><img src="https://agentmods.dev/badge/skills/mirkobozzetto/arsenal/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 | $0.00045 | $0.00982 |
| Opus 5.5 | $0.00018 | $0.00393 |
| Sonnet 5.5 | $0.00009 | $0.00196 |
| Haiku 4.5 | $0.00005 | $0.00098 |
Grade A, and why
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 14d 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 — 96 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Review
Analyze code in depth, say what is wrong or illogical, and route each point to the skill that fixes it. The review never changes project code.
With a docs/review/*/report.md path, skip to step 5 on that report.
1. Interview
One question per message, in plain text and the conversation language. Wait for each answer. Skip a question the request already answers. Offer your suggestion each time.
- What to analyze: the whole project, a folder, a feature, the current branch against its base, or a pull request.
- What to look for: bugs, security, needless complexity, performance, consistency with what was planned (briefs, proposals, issues), another domain the user names, or all of them.
- What worries the user in particular, or nothing.
- Which model and thinking level for the readers. Default: the session's. Offer only what the harness allows (references/harness.md).
- Consent: « Je lance N relecteurs en lecture seule sur : on y va ? ». N is one per chosen angle. This answer is the delegation consent; without it, review solo.
2. Readers
Launch one reader per angle, in parallel, read-only, through the harness
adapter in references/harness.md. Each receives: the scope (paths or diff
command), the angle's brief from references/angles.md, the user's worry,
and the planned work when the angle is consistency. Each returns only a JSON
array of {severity, file, line, issue, evidence, fix_skill}, where
severity is critical, major or minor and fix_skill is ship,
issue, brief or propose.
No reader available, refused consent, or a failed launch: review each angle yourself, one after the other, and say so in one line.
3. Verify
The lead checks every finding before the report:
- open the cited file at the line and confirm the problem is real;
- read what calls that code: with GitNexus, when its index answers, use
contextandimpact; otherwise a plain text search. GitNexus missing: recommend it once in one line, never require it. GitNexus failing: fall back to search without retrying it; - reproduce with an existing command or a disposable check when cheap.
What ships with it
4 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.
- 14d ago First seen · 96 lines · 45 tokens per session scan A 23f7d02e42e6
review is a skill published in the GitHub repository mirkobozzetto/arsenal (20 stars, last pushed 14d ago), licensed MIT. It adds 45 tokens to every session and 982 once invoked, about $0.0002 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 2026-10-02.
Other skills, from other repositories
refactor
Guides a refactor, cleanup, or restructure with the right decomposition. Use when the user asks to refactor, simplify, extract, or modernize code.
pre-publish-review
Nuclear-grade 12-agent pre-publish release gate. Runs /get-unpublished-changes to detect all changes since last npm release, spawns up to 10 ultrabrain agents for deep per-change analysis, invokes /review-work (orchestrator manual QA plus one gate reviewer) for holistic review, and 1 oracle for overall release…
work-with-pr
Full PR lifecycle in a fresh task-owned git worktree: implement via the ulw-loop skill with mandatory evidence-bound manual QA → reviewer-readable English PR → verification loop (CI + Cubic, where Cubic is skipped only when its quota is exhausted) → merge by default → worktree cleanup. Decomposes one task into the…
remove-ai-slops
Removes AI-generated code smells from branch changes or an explicit file list behind regression tests. Use when the user asks to clean up, deslop, or remove AI-slop patterns from recent changes.
review-work
Post-implementation gate review: run manual QA on the real surface yourself, then launch ONE gate reviewer (never a panel) to audit goal, constraints, code quality, security, missed context, and QA evidence. Use before a PR handoff or when the user explicitly asks to review completed work.
tech-debt-audit
Thorough, file-cited technical debt audit across 9 dimensions using AST-grep (tree-sitter), grep, LSP, and language-native tooling. Produces TECHDEBTAUDIT.md with severity, effort estimates, and prioritized fixes. Use when asked for codebase health check, tech debt audit, architecture review, code quality assessment…