running-test-sessions
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill facilitates software testing workflows which include installing dependencies, pulling images, and starting services. It explicitly includes a safety check requiring the agent to explain the impact and obtain user confirmation if these actions exceed existing authorizations.
- [INDIRECT_PROMPT_INJECTION]: The skill identifies several files as data sources, including project specifications (Spec), test charters (TCO), and an artifact routing file (
test-privateer/references/artifact-routing.md). While these files represent an ingestion surface for potentially untrusted data, the skill instructs the agent to record raw observations and interpret results within a structured testing context, mitigating the risk of malicious instructions in those files overriding core agent behavior.
Audit Metadata