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 petekp/claude-code-setup --skill pr-self-reviewgit clone --depth 1 https://github.com/petekp/claude-code-setupWrote 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/petekp/claude-code-setup/pr-self-review)<a href="https://agentmods.dev/skills/petekp/claude-code-setup/pr-self-review"><img src="https://agentmods.dev/badge/skills/petekp/claude-code-setup/pr-self-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/petekp/claude-code-setup/pr-self-review"><img src="https://agentmods.dev/badge/skills/petekp/claude-code-setup/pr-self-review.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00036 | $0.00582 |
| Opus 5.5 | $0.00014 | $0.00233 |
| Sonnet 5.5 | $0.00007 | $0.00116 |
| Haiku 4.5 | $0.00004 | $0.00058 |
Grade A, and why
pr-self-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 7d 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
PR author context
Draft notes that help a teammate understand changed behavior, affected callers, and non-obvious choices. This workflow authorizes inspection and drafting, not additional code changes or replies to teammates.
Here, self-review means inspecting the author's own changes and explaining them to reviewers. Publish that context as standalone comments. Do not create or submit a grouped GitHub review, or issue an approval or request-changes verdict.
For an existing PR, read target and authorship to verify the repository, open PR, author, and current head. When no PR exists, inspect the local diff and continue with private drafts.
Read the diff, existing discussion, and affected consumers. Follow history or linked issues when needed to support a reason. Cover distinct reviewer questions across the meaningful changes; there is no required number or ceiling. A complex change may need several short notes, and a straightforward change may need none.
Use writing notes when choosing or drafting explanations. Keep each note short, concrete, and supported. Do not invent the author's intent or speak as Pete personally. Flag a needed durable code comment separately unless its edit is already authorized.
Distinguish the two comment formats:
- PR conversation comments: ordinary comments in the PR conversation, with links to the relevant committed code. These are not attached to a review.
- Single inline comments: individual notes on changed lines, equivalent to Add single comment, not Start a review. GitHub calls these review comments internally, even when posted individually.
Follow the user's placement preference. If "regular" or "non-review" could mean either format in context, clarify that distinction before posting. Do not silently switch formats.
Show the chosen format and every proposed public word before posting. For inline notes, include each path, side, line, and body. For conversation comments, include the code links in the proposed text. Put supporting evidence outside the proposed comments. Add a general comment only when it explains a useful point spanning files; do not invent a review summary or status announcement.
Wait for approval of the wording, format, and anchors or links. Exact supplied text plus an instruction to post counts as approval; do not ask twice. If those approved details need changing, show the revision for approval first. Read existing comments on each run and avoid duplicates.
After approval, read GitHub posting. Post the approved standalone comments through the matching comments endpoint and verify GitHub stored the intended text and location. Do not post when authorship differs, the PR is closed, or approval is missing. Continue private drafting where possible.
What ships with it
3 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.
- 7d ago Changed · +9 lines · +7 tokens per session 90c29840cdf3
- 1mo ago Changed · -160 lines · -81 tokens per session c1ea75d4862b
- 1mo ago First seen · 181 lines · 110 tokens per session scan A c3a8c8a703cd
pr-self-review is a skill published in the GitHub repository petekp/claude-code-setup (48 stars, last pushed 2d ago), licensed MIT. It adds 36 tokens to every session and 582 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 2026-08-31.
Other skills, from other repositories
python-engineering-guidelines
General Python engineering standards for writing, reviewing, and refactoring maintainable code. Use when Codex works on Python code, especially in medium/large projects, runtime adapters, APIs, SDK integrations, async services, stateful systems, or refactors where readability, explicit typing, module boundaries, side…
code-review
Perform thorough code reviews with security, performance, and maintainability analysis. Use when user asks to review code, check for bugs, or audit a codebase.
swe-review
Review an SWE brief, research report, spec, delivery plan, execution state, implementation diff, branch, or pull request. Use for fresh-context, evidence-backed findings before the human chooses the next action.
requesting-code-review
Pre-commit review: security scan, quality gates, auto-fix.
html-artifacts
Author the HTML for a plan artifact, dashboard iframe, or Slack attachment — structure, design plan, available runtime, theming, and craft. Read this before writing HTML for publishartifact, outputiframe, or slackattachhtml.
baby-sit
Monitor a GitHub pull request until CI is green, diagnose failures, and rerun only evidence-backed flaky GitHub Actions jobs.