neuron-one/GODMODE - Page 2

auth-patterns

Implements secure authentication patterns. JWT with refresh tokens, OAuth 2.0, RBAC, session management. Use when building auth for any project. Don't use for social login setup.

code-security

Security audit for application code. Scans for OWASP Top 10, authentication flaws, data exposure, injection attacks. Don't use for infrastructure or network security.

git-guardrails-claude-code

Installs PreToolUse hook + bash script to hard-block dangerous git commands. One-time setup for settings.json. Use when user wants to install git safety hooks. Don't use for soft rules — use git-guardrails instead.

git-guardrails

Soft rules for Claude to avoid dangerous git commands (force-push, hard reset, branch -D). Always-active behavioral guardrail. Don't use for hook setup — use git-guardrails-external instead.

secrets-guard

Finds exposed secrets, API keys, tokens, and credentials in codebase. Checks .env files, git history, hardcoded values. Don't use for runtime security monitoring.

ship-safe

16 parallel security agents scanning 80+ attack classes. Prompt injection, RAG poisoning, API key leaks, supply chain, MCP misconfigurations. Use before production deployment. Don't use for minor code changes.

supply-chain

Checks npm/pip dependencies for known vulnerabilities, typosquatting, and malicious packages. Use before adding new dependencies or during security audits. Don't use for runtime monitoring.

app-store

Pre-submission check for iOS App Store guidelines. Payments/IAP compliance, privacy manifests, login flows, metadata quality, binary validation. Use before submitting to Apple.

browser-testing

Automated browser testing and visual verification. Opens app in browser, clicks elements, verifies UI renders correctly. Use after frontend changes to verify visually.

e2e-tests

End-to-end testing with Playwright or Cypress. User flows, form submissions, navigation, visual regression. Don't use for unit or API tests.

unit-tests

Unit test writing with Jest (JS/TS) or pytest (Python). Test structure, mocking, edge cases, coverage. Don't use for E2E or integration tests.

dispute

Cross-model dispute with Claude + GPT + Gemini. All three models give independent positions, then see each other's arguments and negotiate through rounds until consensus. Use for stack choices, architecture decisions, or any important technical decision.

interview

Deep interview before building features or projects. Asks 20-50 questions across business, product, technical, security, and marketing blocks. Use before any major feature, new project, or architectural decision.

planning

Creates 3 tracking files (task_plan.md, findings.md, progress.md) and continuously references them throughout execution. Eliminates drift. Use when implementing any feature requiring 3+ steps.

self-improve

Audits skills, CLAUDE.md files, and conversation patterns. Finds redundancies, suggests improvements, identifies learning opportunities. Use periodically to improve GODMODE quality.