project-review
Skill4 starsChanged 6 months agoMITView on GitHub
Use when reviewing modified code, git diffs, commits, merges, PR-like changes, or implementation output.
17 other projects use a version of this.
Adopted by
Most recently taken up by
- loupe-ac/loupetheir own version, first seen Nov 16, 2026
- lwmxiaobei/xiaobei-skillstheir own version, first seen Nov 15, 2026
- Eleven-M/flowmindtheir own version, first seen Nov 8, 2026
- ubermuda/loupetheir own version, first seen Nov 8, 2026
- 70asunflower/70asunflower-skillstheir own version, first seen Nov 8, 2026
- SWEStash/swe-workflow-skillstheir own version, first seen Nov 8, 2026
Other versions17 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 17
- jerry-ai-dev/MODULAR-RAG-MCP-SERVER★ 1.2ksame name
- lwmxiaobei/xiaobei-skills★ 156same name
- jiahao-shao1/sjh-skills★ 43same name
- 5throck/ai-workspace-standards★ 25same name
- rittmananalytics/wire-plugin★ 14same name
- ila/duckdb-claude-skills★ 4same name
- loupe-ac/loupe★ 2same name
- ubermuda/loupe★ 2same name
- 70asunflower/70asunflower-skills★ 1same name
- davidsolheim/skills★ 1same name
7 more not listed
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 repository30 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- .cursor/rules/agent-feed.mdcCursor rule
- concept-reviewSkill
- design-reviewSkill
- docxSkill
- engineering-planningSkill
- frontend-designSkill
- guidance-promoterSkill
- project-architectureSkill
- project-developmentSkill
- project-fixSkill
- skill-creatorSkill
- skill-maintainerSkill
- specialist-routerSkill
- xlsxSkill
- concept-reviewSkill
- design-reviewSkill
- docxSkill
- engineering-planningSkill
- frontend-designSkill
- guidance-promoterSkill
- project-architectureSkill
- project-developmentSkill
- project-fixSkill
- project-reviewSkill
- skill-creatorSkill
- skill-maintainerSkill
- specialist-routerSkill
- xlsxSkill
The file
We split this skill into the rules it gives an agent
6 rules. 0 are common, 4 are distinctive, and 2 are very specific to this project.
0
Common
Most of the 12 edited versions we read keep these.
4
Distinctive
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 12
Use this skill for code, diff, commit, merge, or AI-generated implementation review.
- Distinctive: line 14
Do not modify code during a review unless the user explicitly asks for fixes in the same task.
- Very specific: line 25
If a finding requires an unconfirmed decision outside the current Task Brief, stop and apply .agents/rules/decision-gates.md.
- Very specific: line 45
Use .agents/skills/concept-review/SKILL.md when the diff introduces or changes names, concepts, abstractions, protocol terms, user-facing …
- Distinctive: line 51
Check for evidence of:
- Distinctive: line 64
If a relevant item has no evidence, report it as a finding or residual risk.
“Common” is measured: kept by at least 6 in 10 of the 12 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Project Review Skill
Required Use
Use this skill for code, diff, commit, merge, or AI-generated implementation review.
Do not modify code during a review unless the user explicitly asks for fixes in the same task.
Review Mode
Classify the review before acting:
Pure review: the user asked for review, diff review, commit review, or findings. Do not edit files. Return findings and residual risk only.Implementation gate: the review is part of a task that already includes coding, refactor, fix, or tests. Findings may become follow-up fixes inside the same Task Brief.
For implementation gates, route P0/P1 findings and default-fix P2 findings through .agents/skills/project-fix/SKILL.md, then rerun relevant verification before final handoff.
If a finding requires an unconfirmed decision outside the current Task Brief, stop and apply .agents/rules/decision-gates.md.
Required Reading
.agents/rules/outcome-boundary.md.agents/project/architecture-boundaries.md.agents/rules/development-workflow.md.agents/project/project-structure.md.agents/rules/testing-gates.md.agents/rules/review-gates.md.agents/domain/contracts.md.agents/skills/README.md.agents/skills/specialist-router/SKILL.mdwhen an optional, imported, custom, or specialized skill may match the diff or risk.
Review Scope
Findings should prioritize correctness, contract drift, module ownership, tests, error handling, security/secret handling, and maintainability.
Before finalizing findings, check .agents/skills/README.md for specialized review or fix skills that match the changed language, architecture, framework, persistence layer, or risk category. Use .agents/skills/specialist-router/SKILL.md when the match is not obvious, when a custom skill is involved, or when multiple skills could apply. Use only the skills that directly serve the current review boundary.
Use .agents/skills/concept-review/SKILL.md when the diff introduces or changes names, concepts, abstractions, protocol terms, user-facing language, or skill vocabulary.
Module Understanding Review
For every affected module, command, adapter, protocol asset, or public surface, review whether the change proves the AI understood the path it modified.
Check for evidence of:
- Affected entrypoints, owners, and outputs: the changed files cover the actual requirement path and do not miss a relevant caller, adapter, command, generated template, or user-visible surface.
- Flow closure: input or trigger -> orchestration/state -> output, write, generated asset, or side effect is complete.
- Contract consistency: function signatures, CLI arguments, config keys, schemas, generated file paths, adapter expectations, and documented behavior stay aligned.
- Boundary and failure handling: validation, permissions, error messages, cleanup, rollback expectations, and logging/audit anchors are handled at the owning layer.
- State and data safety: defaults, empty/null cases, async or repeated execution, durable state, secrets, and trust/config boundaries remain safe.
- Reuse and duplication: existing helpers, patterns, tests, adapters, or template owners were reused before adding parallel behavior.
- Maintainability cost: new concepts, abstractions, branches, or files add only necessary complexity.
- Minimal implementation: the diff does not introduce future-facing machinery, broad rewrites, or unrelated formatting/noise to satisfy a narrow task.
- Style consistency: similar existing behavior uses compatible naming, layering, errors, output style, comments, and verification patterns.
- Shared or deletion risk: global files, shared helpers, generated assets, removed paths, and adapter mirrors were checked for other consumers.
If a relevant item has no evidence, report it as a finding or residual risk. Do not replace missing evidence with a general "looks fine" statement.
Severity Model
P0: Data loss, secret leak, source-of-truth break, or irreversible corruption.P1: Public API, contract, lifecycle, persistence, security, or recovery break.P2: Module boundary, test coverage, traceability, or maintainability risk.P3: Naming, local readability, or small cleanup issue.
Output
Return findings first, ordered by severity. If no findings are found, say so explicitly and list residual risks or tests not run.
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.

