Borrowing it
Nothing to install: this file belongs to schmitthub/clawker. 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/schmitthub/clawker/main/.github/skills/code-review/SKILL.mdgit clone --depth 1 https://github.com/schmitthub/clawkerWrote 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/schmitthub/clawker/code-review)<a href="https://agentmods.dev/skills/schmitthub/clawker/code-review"><img src="https://agentmods.dev/badge/skills/schmitthub/clawker/code-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/schmitthub/clawker/code-review"><img src="https://agentmods.dev/badge/skills/schmitthub/clawker/code-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 | $0.00028 | $0.00606 |
| Opus 5.5 | $0.00011 | $0.00242 |
| Sonnet 5.5 | $0.00006 | $0.00121 |
| Haiku 4.5 | $0.00003 | $0.00061 |
Grade A, and why
code-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 1mo 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 — 58 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Follow these steps in order. .github/copilot-instructions.md defines each
priority in full and the comment style; this file is the procedure.
1. Lint suppression
Search every added or changed line in the diff for //nolint, // nolint,
#nosec, and //lint:ignore. Search .golangci.yml changes for
exclusions, exclude, and nolintlint. List every hit as file:line.
Report each hit as its own comment (priority 1).
2. Control plane servers and listeners
Search the diff under controlplane/, internal/controlplane/,
cmd/clawkercp/, and clawkerd/ for grpc.NewServer, net.Listen,
net.ListenConfig, http.Server, and ListenAndServe. For each hit:
- Confirm the gRPC server attaches the
controlplane/authinterceptor in the chain order used bycontrolplane/server/grpc_stack.go. - Confirm the certificate, key, and CA files come from a lane that belongs
to this listener only. Compare against
controlplane/infracerts,controlplane/otelcerts, andcontrolplane/sdscerts. - If the listener uses client certificates instead of bearer tokens, confirm
tls.RequireAndVerifyClientCertand a CN or SAN pin inVerifyPeerCertificate, and confirm the pull request description gives the reason. - Confirm there is no
panic,log.Fatal, oros.Exit, and that each serve goroutine has arecover.
Report each failed check as a blocking comment (priority 2).
3. CLAUDE.md files
For each CLAUDE.md in the diff, read the file mode and the blob content.
Pass only when the mode is 120000 and the content is exactly AGENTS.md,
and a file named AGENTS.md exists in the same directory. Git submodules are
exempt. Report each failure as a blocking comment (priority 3).
4. Version pins
For each added FROM line, uses: line, rev: line, or image constant in
Go source, confirm it carries a @sha256: digest or a full commit SHA.
Report each miss (priority 4).
5. Standing rules
Apply the remaining priority 4 rules and the rule files that
.github/copilot-instructions.md names.
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.
- 1mo ago First seen · 58 lines · 28 tokens per session scan A fb178ca60ffd
code-review is a skill published in the GitHub repository schmitthub/clawker (58 stars, last pushed today), licensed MIT. It adds 28 tokens to every session and 606 once invoked, about $0.0001 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 2026-09-08.
Other skills, from other repositories
prismer-github
GitHub via gh CLI: PRs, issues, reviews, repos, auth.
address-review
Read PR review comments from GitHub, evaluate, address, and reply.
github_read
Read access: repos, issues, PRs, files, branches, workflows, runs.
dev-cycle
Run a requested full development workflow: branch, implement, check, review, PR, merge, and release.
pull-request-guide
Use when authoring a pull or merge request - writing the description, sizing and splitting the change, documenting how it was tested, surfacing tradeoffs, and getting it review-ready before assigning reviewers. Triggers on opening a PR/MR, PR descriptions or templates, large or atomic PRs, what to put in a PR, draft…
pr-review
AI-assisted review of one or more GitHub Pull Requests. Offloads deterministic checks (lint, complexity, duplication, security, perf) to repo linters via self-contained scripts, then applies LLM judgment to what linters cannot catch and posts high-signal findings after a single human approval. Pass one or many PRs…