reproduce
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external data in the form of user-reported symptoms and conditions. An attacker could potentially embed malicious instructions within a bug report to influence the agent during the reproduction phase.
- Ingestion points: User-provided problem descriptions and condition reports (SKILL.md).
- Boundary markers: None specified for the input data.
- Capability inventory: The agent is encouraged to create and run scripts, tests, and browser steps.
- Sanitization: The skill provides a manual rule to remove secrets and personal data from evidence, but no automated sanitization of input strings is mentioned.
- [DYNAMIC_EXECUTION]: The skill's primary function involves creating and running scripts or tests to reproduce problems. This is an intended capability but represents a runtime execution surface.
- [SAFE]: The skill contains explicit safety rules that mitigate risk, including a 'read-only' policy for production environments unless user approval is obtained, and a requirement to scrub PII and secrets from all captured evidence.
Audit Metadata