kenryu42/cc-safety-net
An AI coding agent guardrail — a CLI hook that blocks destructive git and filesystem commands and secret file access before they execute. Supports Amp Code, Antigravity CLI, Claude Code, Codex, Copilot CLI, Cursor, Gemini CLI, Grok Build, Hermes Agent, Kimi Code, OpenClaw, OpenCode, and Pi. observed · 2026-08-28
Health v2 · maintenance only
83/100
- Activity 99
- Release rhythm 99
- Longevity 17
How is this computed?
round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10) — computed 2026-09-03. Adoption (stars, forks) is never an input.
- gap_med: 1
- age_days: 251
- days_rel: 8
- days_push: 8
- n_releases_24m: 40
Adoption not part of the score
1507 stars · 74 forks observed · 2026-08-28
What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded
CC Safety Net is a pre-execution hook for AI coding agent CLIs (Claude Code, Codex, Cursor, Gemini CLI, and many others) that blocks destructive git and filesystem commands and access to secret files before they run. It parses command semantics so bypass tricks like flag reordering, shell wrappers, and interpreter one-liners cannot evade the guard.
Use cases
- stop an AI agent from running rm -rf on my home directory
- prevent claude code from running git reset --hard and losing my work
- block coding agents from reading .env files and SSH keys
- add a safety guardrail to multiple AI coding CLIs
- enforce technical limits on what commands an AI agent can execute
- protect git history from force pushes by AI agents
- keep AI agents from copying AWS credentials
When to choose
- you use AI coding agents that execute shell commands and want hard enforcement, not just prompt instructions
- you want one consistent safety layer across many coding CLIs like Claude Code, Codex, Cursor, and Gemini CLI
- you need to protect secrets such as .env files, SSH keys, and cloud credentials from agent reads and writes
- you want configurable safety levels (Standard, Strict, Paranoid) and custom blocking rules
When to avoid
- you need a true sandbox or OS-level isolation rather than a pre-execution command check
- you use a coding agent CLI not on the supported list
- you want the agent to freely run destructive operations without any blocking
- you need protection against malicious code the agent writes and executes, which command analysis cannot fully catch
Facets
cli-tool · maturity active
security cli developer-tools middleware developer-tools security cli cross-platform windows ai-coding-agents guardrails claude-code pretooluse-hook secret-protection destructive-command-blocking agent-safety codex cursor gemini-cli ai-agents command-line nodejs macos linux
6 sources
- readme: https://github.com/kenryu42/cc-safety-net · fetched 2026-08-28 · 0da609731c5e
- homepage: https://ccsafetynet.com · fetched 2026-08-29 · a4f465451329
- site_page: https://ccsafetynet.com/docs · fetched 2026-08-29 · f0b74125634d
- site_page: https://ccsafetynet.com/docs/reference/blocked-commands · fetched 2026-08-29 · dad418ba0e44
- site_page: https://ccsafetynet.com/docs/reference/secret-protection · fetched 2026-08-29 · 40af698eefbd
- registry_npm: https://registry.npmjs.org/cc-safety-net · fetched 2026-08-29 · 671a06a7d451
Member repositories
| Repository | Role | Health v2 |
|---|---|---|
| kenryu42/cc-safety-net | main | 83 |
For agents
markdown · JSON · MCP: product_card(name="kenryu42/cc-safety-net")
Data as of 2026-08-30T08:39:29.467469+00:00 · Report a problem