forge-code-review
Skill13 starsChanged 44 days agoMITView on GitHub
Independently review an exact code candidate, including UI markup and styles, for reachable defects, regressions, security, engineering standards, and test quality. Use when the user asks for a standalone code review, PR review, branch review, commit review, diff inspection, or code-only critique. Do not use for implementation, repair, full Forge lifecycle Review, acceptance testing, pure visual design artifact review, or knowledge-work review.
2 other projects use a version of this.
Adopted by
Most recently taken up by
- ForgeyClap/claude-forgetheir own version, first seen Oct 11, 2026
- ForgeyClap/Forge-Coretheir own version, first seen Oct 11, 2026
Other versions2 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 2
- ForgeyClap/claude-forge★ 2same name
- ForgeyClap/Forge-Core★ 2same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository9 other files it gives its agents.
- AGENTS.mdAGENTS.md
- forge-buildSkill
- forge-planSkill
- forge-reviewSkill
- forge-shipSkill
- forge-simplifySkill
- forgeSkill
- forge-specSkill
- forge-verifySkill
The file
We split this skill into the rules it gives an agent
6 rules. About 0 are common, 4 are distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈4
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 18
Read the target repository's root and applicable nested instruction maps before judging code.
- Very specific: line 35
When Forge Review assigns this skill, use the exact candidate, base, path scope, authority packet, actual applicable target …
- Distinctive: line 43
Follow the code-review procedure.
- Very specific: line 60
Use PASS, REVISE, RETHINK, READY_FOR_USER, or BLOCKED.
- Distinctive: line 69
Use the concise Code Review report.
- Distinctive: line 84
Every finding has authority, reachability, evidence, consequence, severity, and proportionate remedy
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Forge Code Review
<setup>
Load [authority](../forge-shared/references/authority.md),
[impact-based admission](../forge-shared/references/severity.md),
[target harness resolution](../forge-shared/references/target-harness.md), and
[returns](../forge-shared/references/returns.md); these provide the common judgment
and evidence contract without starting a lifecycle or Launch.
Adopt the bundled [Reviewer](../../agents/reviewer/instructions.md) as the senior
engineering lens. Stay adversarial in investigation, conservative in findings,
and read-only throughout.Read the target repository's root and applicable nested instruction maps before
judging code. Follow only the standards links relevant to the candidate, and
re-resolve that authority when investigation enters another subtree or standards
domain. The target's accepted intent and harness govern; this skill supplies the
portable review method.
</setup>
<activation>
This is a leaf skill. It may run directly for a standalone code review or as the
Code Review dimension assigned by Forge Review. It never invokes Forge Review,
starts a delivery lifecycle, integrates other dimensions, edits the candidate,
runs Verify, or routes repair.Code includes UI markup and styles, even a CSS-only change. Inspect their source correctness and engineering standards; an assigned Design judge owns the rendered visual judgment. Pure design artifacts without code are outside this leaf.
When Forge Review assigns this skill, use the exact candidate, base, path scope,
authority packet, actual applicable target harness/rule/reference/check paths,
allowed commands, and report boundary in the assignment. Read those sources rather
than relying on an assigning owner's summary. When invoked directly, resolve
those inputs with the procedure.
</activation>
<workflow>
Follow the [code-review procedure](references/code-review.md). Pin the complete
scoped diff before forming hypotheses, trace real callers and observable
consequences, inspect tests as code, and use installed or version-matched evidence
for dependency claims.Consume credible supplied proof before running checks. Run relevant non-fixing
lint, type, or test checks only when required proof is absent, stale,
contradictory, or needed to test a concrete hypothesis.
</workflow>
<finding_policy> Admit a finding only when authority, a reachable current trigger, observed evidence or a concrete causal trace, and a material consequence all survive scrutiny. Apply the shared impact-based severity rules. A Spec mismatch is not automatically P0, and severity does not authorize a remedy outside the accepted scope. Missing required evidence is an explicit gap, not an invented runtime defect.
Use PASS, REVISE, RETHINK, READY_FOR_USER, or BLOCKED. PASS means the
Code Review dimension found no unresolved P0/P1 and has sufficient evidence for
its claims. REVISE means supported correction; RETHINK means the mechanism needs
reconsideration; READY_FOR_USER identifies a consequential authority/intent choice;
BLOCKED means required evidence or capability prevents judgment. Missing required
proof precludes PASS. It does not establish integrated Forge Review or Verify.
</finding_policy>
<output>
Use the concise [Code Review report](assets/report.md). Return the selected skill
and source, exact candidate/base and path scope, inspected authority including
actual applicable app standard/reference paths, checks and
observed evidence, findings, gaps, and one Code Review verdict. The record-owning
coordinator preserves the return as its own labelled section
or linked managed document when Forge Review assigned it; this reviewer never
writes records.
</output><checklist>
- Exact candidate, base, dirty state, and scoped paths are reproducible
- Applicable accepted intent and target-repository standards were resolved
- Complete scoped diff preceded hypotheses and focused source exploration
- Real callers, boundaries, async/state/error paths, and test validity were traced as applicable
- Dependency claims use installed source/types or authoritative version-matched evidence
- Every finding has authority, reachability, evidence, consequence, severity, and proportionate remedy
- Public input behavior was not inferred only from a golden fixture
- Report is read-only, dimension-scoped, concise, and honest about gaps
</checklist>Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.