anvil-code-review

Reviews git diffs against coding standards and originating specs using parallel sub-agents.

1|Updated Aug 27, 2026
One-click install
npx skills add https://github.com/rogerznts/anvil --skill anvil-code-review-rogerznts
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: anvil-code-review
Source: https://github.com/rogerznts/anvil/tree/main/anvil/.claude/skills/anvil-code-review
Command: npx skills add https://github.com/rogerznts/anvil --skill anvil-code-review-rogerznts

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Reviewing a branch or PR requires checking two independent questions at once: does the code follow the repo's documented standards, and does it actually implement what the issue or spec asked for. Mixing both in one review lets one axis mask the other. ## Core Features & Use Cases - Two-axis review: Runs a Standards review and a Spec review as parallel sub-agents so their contexts stay separate, then reports findings side by side without merging or reranking. - Fowler smell baseline: Applies a fixed set of code smells from Refactoring ch.3 (Duplicated Code, Feature Envy, Shotgun Surgery, etc.) even when the repo documents no standards, with documented repo rules always overriding the baseline. - Spec traceability: Locates the originating spec via issue references in commit messages, a user-supplied path, or files under docs/, specs/, or .scratch/, and quotes spec lines for each finding. - Use Case: Before merging a feature branch, ask for a review since main; the skill diffs against the merge-base, checks the diff against CODING_STANDARDS.md and the linked issue, and reports violations and missing requirements separately. ## Quick Start Ask the assistant to review the changes since main using the anvil-code-review skill and report standards and spec findings separately.

Frequently Asked Questions about anvil-code-review

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I review a git branch against coding standards and its spec?▼

Provide a fixed point such as a commit SHA, branch, tag, or HEAD~5, and the skill diffs it against HEAD using a three-dot merge-base comparison. It then runs a Standards sub-agent and a Spec sub-agent in parallel and reports both sets of findings side by side.

What happens if my repo has no documented coding standards?▼

The Standards axis falls back to a fixed baseline of Fowler code smells from Refactoring chapter 3, such as Duplicated Code, Feature Envy, and Shotgun Surgery. Each smell is reported as a labelled judgement call, never a hard violation, and documented repo standards always override the baseline.

How does the review find the spec for a branch?▼

It looks for issue references in commit messages first, then a user-supplied path, then spec files under docs/, specs/, or .scratch/ matching the branch or feature. If nothing is found it asks the user, and if no spec exists the Spec axis is skipped and noted in the report.

Why are standards and spec findings reported separately instead of merged?▼

A change can pass one axis and fail the other: code can follow every standard while implementing the wrong thing, or match the spec while breaking conventions. Keeping the reports separate prevents one axis from masking the other, so findings are never merged or reranked across axes.

What are the limitations of this diff review approach?▼

The review requires a resolvable fixed point and a non-empty diff, failing early otherwise. It also depends on the repo's issue tracker configuration in docs/agents/issue-tracker.md, and baseline smell findings are judgement calls rather than enforced violations.