engflow-code-review
Skill0 starsChanged 17 days agoMITView on GitHub
Independently review code changes for correctness, requirement coverage, regressions and validation gaps. Use on explicit review request. No Spec or tracker is required; Implement consumes the shared review method directly.
In this repository13 other files it gives its agents.
- AGENTS.mdAGENTS.md
- diagram-aestheticsSkill
- engflow-clarifySkill
- engflow-designSkill
- engflow-handoffSkill
- engflow-implementSkill
- engflow-planSkill
- engflow-pocSkill
- engflow-refactor-assessSkill
- engflow-researchSkill
- engflow-specifySkill
- engflow-tasksSkill
- semver-bumpSkill
The file
We split this skill into the rules it gives an agent
3 rules. About 0 are common, 2 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 25
Read and apply the shared review contract for review priorities, evidence thresholds, severity/confidence, counterevidence, reviewer …
- Distinctive: line 32
Use fresh-context read-only reviewers when supported and proportionate.
- Very specific: line 62
All reviewers remain read-only.
≈ “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
Engflow Code Review
Provide a standalone, read-only assessment of the requested code change. Return evidence-backed findings in conversation, not a patch, broad refactoring agenda or approval to merge. This entrypoint is for everyday working-tree, staged, commit, branch and PR reviews independent of who produced the code.
Review procedure
Establish the exact requested target, baseline and actual source state. Follow Git scope guidance to distinguish working files, staged content, commits and branch/PR comparisons. Ask only when ambiguity materially changes coverage; do not silently omit relevant untracked files or substitute the checkout for a different target revision.
Read and apply the shared review contract for review priorities, evidence thresholds, severity/confidence, counterevidence, reviewer independence and finding format. Inspect the diff and enough surrounding code, consumers, project rules, requirements and existing test evidence to establish impact. A Spec or tracker is not mandatory; missing artifacts limit coverage but do not block an otherwise useful review.
Use fresh-context read-only reviewers when supported and proportionate. Brief them neutrally with the target, baseline, requirements, source pointers and constraints, not the author's preferred conclusion. Verify consequential returns against the actual source and seek counterevidence. Otherwise review directly and disclose self-review when assessing work authored in the same context.
Findings and limits
Report actionable findings with severity, confidence, precise path/line or diff location, trigger, impact and a minimal correction or verification direction. Rank by impact and consolidate duplicate causes. Separate established findings from material open questions and relevant pre-existing issues. Include the review scope, actual checks, coverage limits and whether review was independent or self-review. The absence of substantiated findings is not proof of correctness, successful tests or security, and does not approve a merge.
Workflow independence and authority
Review suitable changes regardless of which tools or Skills produced them. When available, use effective Spec, Design and Plan material to check implementation coverage and justified deviations. An ambiguous requirement is an open question, not permission to invent intended behavior. This Skill assesses code and its compliance; it does not replace review of whether requirements or the technical proposal themselves are sound.
Within an explicitly requested implementation task, Implement reads the shared review method directly and owns the review/fix/validation lifecycle. That method reuse does not invoke this Skill entrypoint. Other workflow references alone do not invoke review.
All reviewers remain read-only. Inspect files and existing results; do not run mutating builds/tests/experiments, install dependencies, write reports or source, commit, switch branches, manage worktrees or publish comments/approvals. Recommend missing verification to the authorized implementer. Fixes or publication require their own authority.
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.

