review-local

Review local git changes from 8 expert perspectives using multi-agent team orchestration. Produces a consolidated report with Critical/Important/Nice-to-have severity levels. Lightweight pre-commit or pre-push quality gate — no PR or branch push required. Use when the user asks to review local changes, check changes before committing, get a team review of working tree changes, or run a pre-commit review. Trigger phrases include "review local", "review my changes", "review local changes", "pre-commit review", "review before commit", "review before push", "team review my changes", "check my changes", "review working tree", "local code review", "review diff", "review my diff".

Safety Notice

This listing is imported from skills.sh public index metadata. Review upstream SKILL.md and repository scripts before running.

Copy this and send it to your AI assistant to learn

Install skill "review-local" with this command: npx skills add hideki5123/agent-skill-set/hideki5123-agent-skill-set-review-local

Local Review

Review current local git changes from multiple expert perspectives using team orchestration.

Arguments

ArgumentDefaultDescription
--staged-onlyfalseOnly review staged changes (git diff --cached)
--filenoneWrite report to review-local-YYYY-MM-DD-HHmm.md

Workflow

1. Collect Changes ──► 2. Read Context ──► 3. Team Review ──► 4. Synthesize ──► 5. Report

Phase 1: Collect Changes

Gather local changes:

# Staged changes (always collected)
git diff --cached

# Unstaged changes (skip if --staged-only)
git diff

# Untracked files (skip if --staged-only)
git status --porcelain | grep '^??'

If all produce empty results, inform the user: "No local changes detected. Nothing to review." and exit.

For untracked files, read the full file content as a pseudo-diff.

If more than 30 files changed, warn the user and ask whether to proceed or narrow scope.

Phase 2: Read Context

For each changed file, read the full file to give reviewers surrounding context. Skip deleted files and binary files.

Phase 3: Team Review

Create an agent team to explore this from different angles: one teammate on software architecture, one on security, one on infrastructure design, one on QA/QC, one on TDD, one on BDD, one on idempotence, and one playing devil's advocate.

Each teammate receives:

  1. The combined diff from Phase 1
  2. The file context from Phase 2
  3. Their perspective-specific prompt — read from references/perspectives/<perspective>.md

Read references/perspectives/ for each teammate's detailed review prompt, questions, output format, and severity guidance.

Phase 4: Synthesize Findings

After all teammates complete their reviews, synthesize findings. Read references/synthesis-protocol.md for the full protocol:

  1. Collect all findings from each teammate
  2. Categorize by severity: Critical, Important, Nice-to-have
  3. Deduplicate overlapping concerns
  4. Prioritize by risk and effort

Phase 5: Present Report

Present the consolidated report inline. If --file is set, write to review-local-YYYY-MM-DD-HHmm.md in the current directory.

Read references/report-template.md for the exact report format.

Behavior Scenarios

Scenario: Full team review of local changes
  Given the user has local changes (staged, unstaged, or untracked)
  When /review-local is invoked
  Then a team is created with all 8 perspectives, each reviews the diff,
  findings are synthesized into a consolidated severity report inline

Scenario: Staged-only review
  Given the user has staged changes ready to commit
  When /review-local --staged-only is invoked
  Then only the staged diff is reviewed by the team

Scenario: Write report to file
  Given the user wants a persistent record
  When /review-local --file is invoked
  Then the report is written to review-local-YYYY-MM-DD-HHmm.md
  and the file path is confirmed to the user

Scenario: No local changes
  Given the working tree is clean
  When /review-local is invoked
  Then the user is told "No local changes detected" and the skill exits

Scenario: Large changeset warning
  Given more than 30 files are changed
  When /review-local is invoked
  Then the user is warned and asked whether to proceed or narrow scope

References

  • references/perspectives/ — One file per review perspective with prompts, questions, output format, and severity guidance
  • references/synthesis-protocol.md — How to merge, deduplicate, and categorize findings
  • references/report-template.md — Exact markdown format for the consolidated report

Source Transparency

This detail page is rendered from real SKILL.md content. Trust labels are metadata-based hints, not a safety guarantee.

Related Skills

Related by shared tags or category signals.

Coding

openclaw-version-monitor

监控 OpenClaw GitHub 版本更新,获取最新版本发布说明,翻译成中文, 并推送到 Telegram 和 Feishu。用于:(1) 定时检查版本更新 (2) 推送版本更新通知 (3) 生成中文版发布说明

Archived SourceRecently Updated
Coding

ask-claude

Delegate a task to Claude Code CLI and immediately report the result back in chat. Supports persistent sessions with full context memory. Safe execution: no data exfiltration, no external calls, file operations confined to workspace. Use when the user asks to run Claude, delegate a coding task, continue a previous Claude session, or any task benefiting from Claude Code's tools (file editing, code analysis, bash, etc.).

Archived SourceRecently Updated
Coding

ai-dating

This skill enables dating and matchmaking workflows. Use it when a user asks to make friends, find a partner, run matchmaking, or provide dating preferences/profile updates. The skill should execute `dating-cli` commands to complete profile setup, task creation/update, match checking, contact reveal, and review.

Archived SourceRecently Updated
Coding

clawhub-rate-limited-publisher

Queue and publish local skills to ClawHub with a strict 5-per-hour cap using the local clawhub CLI and host scheduler.

Archived SourceRecently Updated