Product Launchagentssecurity guardrailsgit protectionplatform integration
Leash Protects AI Coding Agents From Destructive Commands
8.0
Leash, a new CLI tool, sandboxes AI coding agents' file-system access and blocks dangerous commands such as rm, destructive git operations, and symlink escapes, while protecting .env and .git files. It provides platform hooks for OpenCode, Pi, Claude Code and Factory Droid, installs via npm, offers near-zero latency, and enables safer agent execution in developer workflows.
Key Points
- 1Sandboxes agent file-system access and blocks dangerous commands, symlink escapes, and destructive git operations
- 2Prevents agent hallucinations from deleting files or rewriting repository history across multiple platforms
- 3Enables developers to run coding agents safely via npm installable hooks with near-zero latency
Scoring Rationale
Practical agent-security tool with clear usability and official packaging; limited groundbreaking novelty beyond niche agent safety.
Sources
Public references used for this report.
Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems
