Skill Review

The review walk runs in a fresh subagent per iteration (Pattern A — same shape as verify-diff and publicity-review); Edit application stays in the main thread to keep the reviewer bias-free. The skill loops the dispatch + apply cycle until the subagent returns no more mechanical_edits, max iterations is reached, or a safety rail trips. Designed to be called from non-interactive routines such as dev-workflow-triage (d2) or dev-workflow hooks.on_complete; it never prompts the user.

Invocation contract

The caller passes these fields in natural language (the skill extracts them from the invocation text):

• Base ref (optional, default <working-tree-vs-HEAD>) — git ref to diff against. When omitted, the skill looks at the working tree's uncommitted + staged changes (the default scope for dev-workflow post-implementation review). When specified (e.g. Base ref: main), the skill switches to git diff <Base ref> semantics — useful for callers like dev-workflow-triage that want to review a stack of already-committed changes between a base branch and HEAD.
• Max iterations (optional, default 3) — upper bound on the refinement loop. Default 3 mirrors verify-diff's default; prose-quality polish typically converges in 1–2 iterations.

The caller must not stage changes while this skill is running. The skill reads the working tree; staged content would mix into the diff and corrupt the verdict. (The Base ref mode reads committed history vs the ref, so staging interference applies only to the default working-tree mode.)

Process

Step 1 — Detect changed skill files (main thread)