agentskills.codes
RE

release-check

Comprehensive release readiness audit covering security, dependencies, community files, and code quality. Use when asked for a "release check", "release readiness" review, or "pre-release verification".

Install

mkdir -p .claude/skills/release-check && curl -L -o skill.zip "https://agentskills.codes/api/skills/download/13234" && unzip -o skill.zip -d .claude/skills/release-check && rm skill.zip

Installs to .claude/skills/release-check

Activation

This is the description your AI agent reads to decide when to run this skill — the better it matches your request, the more reliably it fires.

Comprehensive release readiness audit covering security, dependencies, community files, and code quality. Use when asked for a "release check", "release readiness" review, or "pre-release verification".
202 chars✓ has a “when” trigger

About this skill

Release Readiness Audit

Pre-release cross-cutting check: security, dependencies, required community files, and code quality. Documentation and metadata checks live in /release-docs-check — run that separately.


1. Dependency vulnerabilities

npm audit --audit-level=high
  • Confirm zero High / Critical
  • If any remain, distinguish direct vs. transitive in the report
  • Note whether npm audit fix can resolve them

2. Required community files

Use Glob to confirm each file exists:

FilePurpose
LICENSELicense (required)
README.mdProject overview (required)
CONTRIBUTING.mdContribution guide (required)
SECURITY.mdVulnerability reporting policy (required)
CODE_OF_CONDUCT.mdCode of conduct (recommended)
.env.exampleEnvironment variable template (required)

3. Personal / project-specific info leaks

Grep (excluding node_modules, .next, .git):

  • Personal domains / URLs: scan source for \.info, \.dev, \.me patterns
  • Personal names / handles: repository owner names must not be hardcoded in source
  • Project-specific names: your own project or brand names (e.g., my-blog-name) must not remain in source

Allowed locations: package.json, README.md, CONTRIBUTING.md, SECURITY.md, LICENSE, _release/.


4. Secrets / credential leaks

Grep for these patterns:

  • API key prefixes: sk-, AIza, ghp_, xoxb-, AKIA
  • Hardcoded passwords: password\s*=\s*["'][^"']{8,}
  • Bearer tokens: Bearer [A-Za-z0-9+/]{20,}

Read .env.example and verify it contains placeholders only.


5. Code quality

  • Stray console.log: Grep under app/(admin)/baan-admin/api/
  • TODO/FIXME/HACK: Grep under app/, lib/ (minor ones are warnings)
  • Japanese error messages: confirm API responses under app/(admin)/baan-admin/api/ do not contain Japanese (exclude translation files like _lib/i18n.ts)

6. Baseline security configuration

  • .gitignore: .env*, data/, *.pem, .session-secret, etc. are excluded
  • package.json: "private": true is set correctly (appropriate for an application)
  • TypeScript strict: tsconfig.json has strict: true

7. GitHub repository settings (manual)

Cannot be verified from code — confirm manually before release:

  • GitHub Private Vulnerability Reporting enabled (Security tab)
  • Default branch is main
  • Branch protection rules set up (if required)

Output format

Clean: ✅ OK Issue: ❌ description (file:line) Warning: ⚠️ description Manual check: 👤 description

Summary format

## Release Readiness

| Category | Status |
|----------|--------|
| Dependency vulnerabilities | ✅ / ❌ |
| Required community files | ✅ / ❌ |
| Personal info leaks | ✅ / ❌ |
| Secret leaks | ✅ / ❌ |
| Code quality | ✅ / ⚠️ / ❌ |
| Baseline security | ✅ / ❌ |

### Verdict
🟢 READY / 🟡 CAUTION / 🔴 NOT READY

Search skills

Search the agent skills registry