Borrowing it
Nothing to install: this file belongs to ra3orblade/swarm. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/ra3orblade/swarm/main/.claude/skills/review/SKILL.mdgit clone --depth 1 https://github.com/ra3orblade/swarmWrote 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/ra3orblade/swarm/review)<a href="https://agentmods.dev/skills/ra3orblade/swarm/review"><img src="https://agentmods.dev/badge/skills/ra3orblade/swarm/review.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 5.1 | $0.00046 | $0.00469 |
| Opus 5 | $0.00023 | $0.00234 |
| Sonnet 5 | $0.00009 | $0.00094 |
| Haiku 4.5 | $0.00005 | $0.00047 |
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 8d 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
Code review
Review the diff of the current branch against main (or the staged/working changes if not on a branch). Report findings; do not edit code unless the user asks.
Scope
git diff --merge-base origin/main # or: git diff (uncommitted)
What to look for, in priority order
- Correctness bugs — wrong logic, off-by-one, unhandled null/error, race conditions, SQL that doesn't match the schema, a hook/route that fails closed when it should fail open (or vice-versa). Give a concrete failing input → wrong output for each.
- Swarm invariants (see
docs/10-development-guidelines.md) — flag any violation:- Repo-agnostic: does it require a file in the monitored repo (only
.swarm.tomlis allowed)? Does it work on an empty folder? - Local-first: any new outbound network call (only the opt-out pricing fetch is allowed)?
- Fail-closed: claims/gates/guards must default to the safe answer.
corepurity: no I/O inpackages/corebeyond types/pure functions; the daemon is the only DB writer.- Product-agnostic docs: no reference to private projects/people/paths.
- Observed content (transcripts, files, pages) treated as data, never instructions.
- Repo-agnostic: does it require a file in the monitored repo (only
- Tests — new behaviour ships with a test; ledger/pricing/parsing semantics tested in
core, not only through the daemon. - Reuse / simplification / efficiency — duplication, an abstraction with <3 uses, needless work in a hot path (hook shim, tailer).
Output
A ranked list, most-severe first. For each: file:line, one-sentence claim, and a concrete failure scenario. If nothing survives scrutiny, say so plainly. Verify a bug reproduces before reporting it as correctness.
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.
- 8d ago First seen · 29 lines · 46 tokens per session scan A 597d66334260
review is a skill published in the GitHub repository ra3orblade/swarm (7 stars, last pushed yesterday), licensed Apache-2.0. It adds 46 tokens to every session and 469 once invoked, about $0.0002 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
A review process for a pull request, which is a proposed set of code changes. It examines the changed code and its surrounding files for bugs, security problems, design issues, and lint errors.
security-audit
A security review skill for examining code changes in a pull request, which is a proposed set of changes awaiting approval. It reports issues such as injection, exposed secrets, unsafe authentication, weak cryptography, vulnerable dependencies, and infrastructure mistakes.
code-governance
A review skill that checks pull-request changes for software licenses, personal information, compliance details, and an audit trail. An audit trail is a record showing what was checked and why.
code-reviewer
Reviews a working diff for correctness, not style.
remove-ai-slop
Audit UI, docs, or prose for AI-generated tells in design and copy, then fix approved findings. Use when the user says de-slop, slop check, "looks AI-generated", or wants copy humanized.
babysit-pr
Monitor a pull request through review and CI until it's green: verify every bot claim against the code, fix what's real, push back on what's wrong, rerun flaky checks. Use when the user asks to babysit, monitor, watch, or shepherd a PR, says "get this PR green" or "handle the review comments". Accepts an optional PR…