AgentGuard
What it is
AgentGuard is an open-source TypeScript CLI framework for regression-testing AI agents using deterministic rules and provider execution. It offers commands for local testing, environment health checks, and end-to-end behavioral scans that generate CI-ready reports.
The problem it solves
AI agents lack deterministic regression testing tools, making it difficult to verify if prompt or model changes have broken agent behavior or introduced security risks like secret leakage.
What you could use it for
- Validating agent responses against strict output schemas and forbidden content rules in CI pipelines
- Performing end-to-end behavioral scans of agents hosted behind HTTP APIs
- Generating automated test scenarios from system prompts and knowledge bases
- Monitoring agent consistency after changes to prompts, models, or tools
Read the source ↗Back to the library →
Summary written by an AI model from the source; check the source before you rely on it.