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 ayberkcansever/ai-skills --skill review-buildgit clone --depth 1 https://github.com/ayberkcansever/ai-skillsWrote 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/ayberkcansever/ai-skills/review-build)<a href="https://agentmods.dev/skills/ayberkcansever/ai-skills/review-build"><img src="https://agentmods.dev/badge/skills/ayberkcansever/ai-skills/review-build/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/ayberkcansever/ai-skills/review-build"><img src="https://agentmods.dev/badge/skills/ayberkcansever/ai-skills/review-build.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 | $0.00040 | $0.01526 |
| Opus 5.5 | $0.00016 | $0.00610 |
| Sonnet 5.5 | $0.00008 | $0.00305 |
| Haiku 4.5 | $0.00004 | $0.00153 |
Grade A, and why
review-build 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 — 123 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Review Build
Input: $ARGUMENTS
Reads: the branch diff, the spec, the plan (including earlier ## Review
cycles), docs/quirks.md, the repo's AGENTS.md / CLAUDE.md.
Writes: nothing — it returns a verdict and findings; the invoking session
records them.
Setup
- Independence: you run in a fresh, read-only context —
/review-buildforks into aPlansubagent (frontmatter), and implement-plan launches a freshPlansubagent on this file. You see no implementer conversation and never edit. The same model is fine; independence comes from the context. - Model: the Agent tool's strongest offered model, unless pinned here:
review-model:(unset). A pinned model that cannot launch → the launcher asks the user, never substitutes silently. - Plan: the spec and plan absolute paths given; else
<root>/docs/plans/<plan>/(<root>= the main checkout,dirname "$(git rev-parse --path-format=absolute --git-common-dir)") for the plan named in the input, else the current branch name or its ticket key. A branch that looks plan-driven but matches no folder → list<root>/docs/plans/and stop, asking to be re-run with a name. - Venue: the plan's worktree,
<root>/.worktrees/<plan>(by path, not branch — the branch may carry a slug), when a plan resolved; otherwise the current checkout. Run every git command there. - Clean tree:
git status --porcelainmust be empty — uncommitted changes escape the diff. Dirty → stop and report. - Diff: base =
git symbolic-ref refs/remotes/origin/HEAD(elsemain); reviewgit diff <base>...HEAD. - Context: the spec and plan,
docs/quirks.md, and the repo'sAGENTS.md/CLAUDE.mdwhen present. Earlier## Reviewcycles: confirm each priorRfinding's fix removed its cause. - Re-run guard: last
## Reviewentry already saysVerdict: shipat this HEAD (or only docs-only commits since) and no spec decision changed since → report "already reviewed" and stop. - Standalone (no
## Reviewline for this HEAD): run the full suite yourself; it stands in for that line in theshiprule. With a plan, end the report withRecord under ## Review in <plan.md absolute path>:and the cycle line plus findings to append — the invoking session writes them.
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 · 123 lines · 40 tokens per session scan A 5f2892145e55
review-build is a skill published in the GitHub repository ayberkcansever/ai-skills (2 stars, last pushed 14d ago), licensed MIT. It adds 40 tokens to every session and 1,526 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 2026-10-03.
Other skills, from other repositories
code-review
Use when asked to review a PR, MR, branch, or diff, audit changed files, or check code quality.
pr-review
Use when opening, reviewing, or merging a pull request. Covers pre-PR checks (linting, tests, branch naming), PR description requirements, and post-merge cleanup. Not for code review commentary or architecture decisions — use the relevant role guidelines for those.
google-feed-ops
Google Merchant Center (GMC) launch and operation plus the retail data layer: new-account launch sequence (site trust → ToS → verify+claim → business info → shipping/tax → data source → review → programs → Ads link → first campaign), Merchant API v1 mechanics and the gmcops CLI, feed spec and attributes, feed rules…
pn-best-of-n
Run 2–3 competing implementations of the same spec in isolated worktrees, eliminate failures via objective gates, then judge survivors with a separate premium-tier pass. Use for ambiguous algorithm/API/refactor choices with strong tests — not auth/security paths.
pr-review
Review a GitHub Pull Request for bugs, security, performance, and code quality. Use when user asks to review a PR or wants pull request feedback. Don't use for reviewing local uncommitted changes, creating new PRs, or merging branches.
pn-skeptic-challenge
Question the proposed approach and verify it is the best possible way. List alternatives and tradeoffs; suggest plan changes if a simpler or better path exists. Gate on confirmation before proceeding.