aliatx2017/reasonix-hermes/simplify-code

Trust tier 3
  • community
Source check pending

Checking that the SKILL.md file is still in the repository.

Parallel 3-agent cleanup of recent code changes — one reviews correctness, one reviews style/patterns, one reviews security. Merge findings into a unified fix pass.

Install

npx vskill@latest install aliatx2017/reasonix-hermes/simplify-code

SKILL.md

Simplify Code — Parallel Review & Cleanup

When to Use

  • After a multi-step feature or bug fix, before committing
  • When a PR has grown messy across several commits
  • User says "clean this up" or "simplify"
  • Before upstream merge — review our changes against upstream code

Process

Launch three subagent reviews in parallel, then merge findings:

Agent 1: Correctness Review
Review the current branch diff. Focus on:
- Logic errors, off-by-one, nil dereference potential
- Missing error handling
- Race conditions (check mutex usage, goroutine lifecycle)
- Breaking changes to interfaces or public APIs
Flag severity: critical / warning / note
Agent 2: Style & Patterns Review
Review the current branch diff. Focus on:
- Go idiom violations (use go vet findings)
- Inconsistent naming, formatting, comment style
- Over-engineered code — could this be simpler?
- Missing tests for new code paths
- Duplicated logic that could be extracted
Flag severity: should-fix / nice-to-have
Agent 3: Security Review
Review the current branch diff. Focus on:
- Injection vectors (shell, SQL, template)
- AuthZ/AuthN bypass potential
- Secrets exposure (api keys, tokens in logs)
- Input validation gaps
- Path traversal, SSRF
Flag severity: critical / warning / note

Merge & Fix

  1. Collect all three review outputs
  2. Triage: fix criticals first, then warnings, then notes
  3. Apply fixes in priority order
  4. Re-run go build ./... && go vet ./... && go test ./...
  5. Confirm all three agents would now pass

Parallel Dispatch

bash
# The review, security_review, and explore tools can run concurrently
# since they're read-only. Launch all three, wait for all, then fix.

When NOT to Use

  • Trivial single-file changes (one review is enough)
  • Changes already covered by CI
  • Emergency hotfixes (defer cleanup to follow-up)

Related

  • Built-in tool: review — single-agent code review subagent
  • Built-in tool: security_review — single-agent security review subagent
  • Constitution rule: go-vet-clean — go build ./... && go vet ./... must pass
  • Project skill: ready-means-tested — verify fixes with evidence

Security

  • Security scanPassed, 95 of 100Pattern checks on the skill's files.
  • Model reviewPassed

Quality evaluation

  • Task evaluationNot runNo evaluation yet. Run one in Skill Studio.

Source

Popularity

7-day trend score
5

Not refreshed yet.

Works with

Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.

  • Claude Code
  • Cursor
  • GitHub Copilot
  • Windsurf
  • Codex
  • Gemini CLI
  • Cline
  • Amp
  • Roo Code
  • Goose
  • Continue
  • Kilo Code
  • Junie
  • Trae
  • OpenHands
  • Qwen Code
  • AI Agent
  • Tabnine Chat
  • Sourcegraph Cody
  • Amazon Q
  • Firebase Genkit
  • Sweep
  • Devin
  • Phind
  • Mentat
  • Claude Desktop
  • Aider
  • GPT Pilot
  • Plandex
  • bolt.new
  • v0
  • Replit Agent
  • Anse
  • TypingMind
  • BetterChatGPT
  • LibreChat
  • Warp AI
  • Pieces
  • Fig AI
Embed badge
vskill VERIFIED badge
[![vskill VERIFIED](https://verified-skill.com/api/v1/skills/aliatx2017/reasonix-hermes/simplify-code/badge)](https://verified-skill.com/skills/aliatx2017/reasonix-hermes/simplify-code)

Something wrong with this skill or its page? Tell us.