jzOcb/agent-guardrails

Mechanical enforcement tools to prevent AI agents from bypassing established project standards.

26
/ 100
Experimental

This project helps engineering and MLOps teams prevent AI coding agents from bypassing established development standards and introducing critical errors. It takes your project's code and existing rules, then enforces them through pre-commit hooks and validation scripts. This ensures the AI agent adheres to proper imports, avoids duplicating logic, and doesn't expose sensitive information.

Use this if you are using AI coding agents (like Claude Code, Clawdbot, or Cursor) and need to mechanically enforce coding standards, prevent security vulnerabilities, and ensure agent-generated code integrates correctly into your existing systems.

Not ideal if your primary concern is managing natural language prompts or if you don't use AI agents for code generation and modification.

AI-safety software-development DevSecOps code-quality security-auditing
No Package No Dependents
Maintenance 10 / 25
Adoption 5 / 25
Maturity 11 / 25
Community 0 / 25

How are scores calculated?

Stars

10

Forks

Language

Shell

License

MIT

Last pushed

Feb 06, 2026

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/agents/jzOcb/agent-guardrails"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.