A static review of the code currently changed in a project. It traces the affected code paths and reports concrete correctness, security, or specification problems with file and line evidence.

Use it to review an active diff, investigate bug risks, and produce prioritised P1, P2, or P3 findings.

It catches real bugs that may be hidden by looking only at diff snippets or by relying on tests that miss actual runtime behaviour.

review — Skill for Claude Code by heliohq

Install

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.

Any agent

npx skills add heliohq/ship --skill review

Clone the repo

git clone --depth 1 https://github.com/heliohq/ship

Made for: Claude Code.

Or install ship, the plugin that ships this one along with the rest of its 11 skills, 3 hooks, 1 MCP server.

Wrote 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.

README.md

[![agentmods](https://agentmods.dev/badge/skills/heliohq/ship/review.svg)](https://agentmods.dev/skills/heliohq/ship/review)

Your own site

<a href="https://agentmods.dev/skills/heliohq/ship/review"><img src="https://agentmods.dev/badge/skills/heliohq/ship/review.svg" alt="Measured on agentmods" height="20"></a>

Per session 47 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.

When invoked 1,763 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe. Third-party audits

NVIDIA SkillSpector pass 7 Sept 2026

How audits are shown

Origin original No closer match found in the catalogue.

Token cost

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	$0.00047	$0.01763
Opus 5	$0.00023	$0.00881
Sonnet 5	$0.00009	$0.00353
Haiku 4.5	$0.00005	$0.00176

Measured 8d ago against content hash 770c10450642, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-08, from the pricing page.

Security

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.

skills/review/SKILL.md · 243 lines

How it starts

The opening of the file, as written. The whole thing — 243 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Ship: Review

This file is an operating contract for an AI reviewer.

Mission

Find real bugs in the active change scope.
Report findings first, ordered P1, then P2, then P3.
Add a short diagnosis only if multiple findings share one root cause.

Red Flag

Never:

Report a bug before understanding the changed code path
Read only diff hunks — read full changed files
Report a concern without file:line and trigger
Report style nits, refactor wishes, or use B1/B2 severity
Ignore staged or unstaged work in standalone mode
Lead with philosophy instead of findings
Force a diagnosis when findings don't share one root cause
Write a vague "looks good" report with no evidence trail

Valid Findings

Report only issues that meet at least one of these:

violates the spec or acceptance criteria
causes broken behavior or a runtime error
causes data loss, partial writes, or inconsistent state
creates a security or trust-boundary vulnerability
breaks callers, consumers, or shared interfaces after a change
lets tests pass while the real behavior is still wrong

A finding without a traced code path or concrete observation is not a valid finding.

Do not report:

style preferences
naming opinions
speculative future concerns
unrelated refactor ideas
missing comments

Severity

Label	Use when
`P1`	ship-stopping correctness failure, security issue, data loss, or major regression
`P2`	real bug or spec deviation with narrower scope or blast radius
`P3`	concrete lower-impact bug or edge-case failure

P3 is still a real bug. It needs the same evidence standard as P1 and P2.

Context

Use the smallest possible setup contract:

spec: caller-provided, else <task_dir>/plan/spec.md if it exists
task_dir: caller-provided, else .ship/tasks/ad-hoc-review-<branch>
scope: the active change scope = <base>...HEAD plus any staged or unstaged worktree changes

If there is no spec, do a diff-only review and say so explicitly. If there are no changes, write a short clean report and stop.

Read the full file on GitHub · 243 lines

Changes

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 · 243 lines · 47 tokens per session scan A 770c10450642

Subscribe to this mod's changes

review is a skill published in the GitHub repository heliohq/ship (92 stars, last pushed 2mo ago), licensed MIT. It adds 47 tokens to every session and 1,763 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 2026-08-30.

Code review Debugging claude-code-plugin codex-skills cursor

fec-debug-framework

A step-by-step method for finding the cause of front-end build failures, browser errors, broken screens, failed requests, or incorrect data. It separates build, runtime, visual, and API problems and uses evidence to test possible causes.

bovinphang/frontend-craft · 66 tokens

fec-legacy-to-modern-migration

A guide for moving older front-end code—such as JavaScript, jQuery, HTML/CSS, server-rendered pages, or old frameworks—to newer tools while keeping the same behavior. It covers gradual migrations as well as modernising parts of a multi-page site.

bovinphang/frontend-craft · 97 tokens

doubt-driven-development

Subjects every non-trivial decision to a fresh-context adversarial review before it stands. Use when correctness matters more than speed, when working in unfamiliar code, when stakes are high (production, security-sensitive logic, irreversible operations), or any time a confident output would be cheaper to verify now…

addyosmani/agent-skills · 67 tokens

review