| Version | Supported |
|---|---|
| 0.1.x | Yes |
Do not open a public GitHub issue for security vulnerabilities. Use GitHub private vulnerability reporting if available, or contact the repository owner through GitHub.
Please include:
- A description of the issue and potential impact.
- Steps to reproduce or a proof of concept.
- Whether the loop can write files, send messages, call external systems, or touch production-like resources.
Agentic loops can become risky when they are allowed to act repeatedly. This repo keeps loops safe by default:
- Loops must be bounded by
maxAttempts,timeoutMs, or an explicit done verdict. - Risky actions require approval via
requiresApproval: true. - Example loops should be read-only unless write behavior is obvious from the name and README.
- Generated state and logs are local-only and ignored by git.
- Secrets belong in environment variables or a local
.envfile that is never committed.
This repo provides local safety primitives, not a sandbox. A loop runs with the same filesystem and network permissions as the user who starts it. Review any new loop before running it against private data, external services, or production-like systems.