jimweller/review-quick

<!-- markdownlint-disable-file MD041 -->

STARTER_CHARACTER = ⚡

Quick Review

Review work-in-progress and recent commits across 8 focus areas in parallel. No arguments. Auto-detects deltas using the same git conventions as /llmdocs and /readme.

Step 1: Detect Deltas

Compute what changed since the project root was last touched. Span committed, staged, unstaged, and untracked work.

PROJECT_ROOT=$(git rev-parse --show-toplevel)
BASELINE=$(git log -1 --format=%H -- "$PROJECT_ROOT")
[ -z "$BASELINE" ] && BASELINE=$(git rev-list --max-parents=0 HEAD)

git log --oneline ${BASELINE}..HEAD
git diff ${BASELINE} --find-renames --stat
git diff ${BASELINE} --find-renames
git status --porcelain
git ls-files --others --exclude-standard

The diff has no ..HEAD so it spans the baseline through the working tree, including staged and unstaged edits. --find-renames surfaces renames as R<score> entries with both old and new paths so a pure rename does not collapse into an empty diff. git status --porcelain is the authoritative list of every modified, added, deleted, renamed, copied, and untracked entry; use it to confirm rename detection. ls-files --others surfaces new untracked files. Files matching .gitignore are excluded.

Stop only if all four signals are empty: git diff empty, git status --porcelain empty, ls-files --others empty, and git log ${BASELINE}..HEAD empty. Then report "No deltas to review" and stop.

Step 2: Read Untracked Files

For each file from git ls-files --others --exclude-standard, read the full file content. Untracked files have no diff representation.

Step 3: Dispatch 8 Review Agents in Parallel

Send a single message with 8 Task tool uses, one per review area. Each Task uses the corresponding subagent_type:

| Agent | Subagent type | | --------------------- | --------------------- | | Architecture & Design | review-architecture | | Correctness & Bugs | review-correctness | | Operational Readiness | review-ops | | Performance | review-performance | | Code Quality | review-quality | | Security | review-security | | SOLID Principles | review-solid | | Testing | review-testing |

Each Task prompt contains:

1. The full diff output from Step 1
2. The contents of any untracked files from Step 2
3. The instruction: "Review the following changes for your focus area. Return findings as your response."

Do not pack repomix. Each agent has all the context needed in the prompt.

Step 4: Consolidate Findings

After all 8 agents return, produce a single consolidated report:

# Quick Review Report

## Summary

- Total findings: <N>
- High: <count>
- Medium: <count>
- Low: <count>

## Findings by Severity

### High

[All High findings from all 8 agents, ordered by file:line]

### Medium

[All Medium findings from all 8 agents, ordered by file:line]

### Low

[All Low findings from all 8 agents, ordered by file:line]

When to Use

This skill auto-loads when:

• A bug fix or feature implementation has just been completed
• Work is about to be committed or a PR is about to be opened
• A subagent-driven-development task has just passed spec compliance review

It is also slash-invokable for ad-hoc review of current work.