12 slash commands, 4 subagents, and 3 CLAUDE.md templates for real engineering workflows — code review, testing, security, performance, migrations, incident response, and onboarding. Not affiliated with or endorsed by Anthropic; built for use with Claude Code.
Everything in this repo is free to clone and use right now — git clone or download the zip. If it saves you time, a suggested $12 in crypto is appreciated (no payment account is needed; native coins work on Ethereum, Base, Arbitrum, Optimism, Polygon, or BNB, and USDC works on Ethereum, Base, Arbitrum, Optimism, or Polygon):
Power Pack Pro details — bonus contents, payment routes, and automated claim steps.
dacode-dev.github.io — QR code + copy-paste address.
Or directly: 0xc4e8021CdFf1a11946Ed16bd264f77D6B3C0C0e9
No payment login, license key, or tracking. Pay if and when it's useful to you. A GitHub account is required only if you want to claim the private Pro repository.
Sponsors get 3 bonus commands/agents (/changelog-gen, /release-checklist, an architecture-reviewer subagent) in a private companion repo. After sending any amount, open a "Claim Pro access" issue with the chain and transaction hash — a GitHub Action checks it on-chain automatically (usually within a couple minutes) and invites your GitHub account as a collaborator. No email or manual step on my end.
Every command and agent here follows the same rules: concrete file:line findings over vague advice, evidence over guessing, and no padding sections with "N/A" just to look thorough. These are the same standards a careful senior engineer would hold a review to.
Project-level (recommended — scoped to one repo):
./install.sh /path/to/your/projectOr manually, copy what you want:
cp commands/*.md ~/.claude/commands/ # or .claude/commands/ inside a project
cp agents/*.md ~/.claude/agents/ # or .claude/agents/ inside a projectThen pick a CLAUDE.md template from claude-md-templates/ matching your stack, copy it to your repo root as CLAUDE.md, and fill in the bracketed sections.
| Command | Use it for |
|---|---|
/code-review |
Deep review of the current diff — correctness, security, performance, test coverage |
/test-gen |
Generate missing tests for a file or function, run and verify them |
/refactor-safe |
Refactor with behavior-preservation as a hard constraint |
/pr-description |
Write a PR description reviewers actually want to read |
/bug-repro |
Turn a vague bug report into a minimal, confirmed repro |
/security-audit |
OWASP-style pass over a diff or path |
/perf-audit |
Find real performance hotspots, not theoretical ones |
/dependency-audit |
Audit dependencies for CVEs, staleness, and bloat |
/onboarding-doc |
Generate an onboarding doc for an unfamiliar codebase |
/incident-postmortem |
Structured, blameless incident postmortem |
/api-docs |
Accurate API reference generated from actual code, not stale docstrings |
/migration-plan |
Plan a DB/schema migration that won't take down production |
| Agent | Role |
|---|---|
code-reviewer |
Read-only reviewer for diffs/branches — correctness, security, performance |
test-writer |
Writes and runs tests until they pass, reports coverage gaps honestly |
security-auditor |
OWASP-style security pass, exploitability-ranked findings |
refactor-planner |
Plans safe, incremental refactor steps before any code is touched |
Node/TypeScript, Python, and Go — structured starting points covering commands, conventions, architecture, and the "things that will bite you" section that actually matters most.
Every file here is plain markdown. Edit freely — these are a strong starting point tuned for real engineering rigor, not a black box. If a command's checklist doesn't match how your team works, cut what doesn't apply.
Questions or issues: open a GitHub issue.