triage
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and act upon external, potentially untrusted data such as test failure logs and YAML files which may contain instructions targeting the agent.
- Ingestion points: The skill reads failure outputs, YAML test files, and project specifications from the local environment (SKILL.md).
- Boundary markers: There are no explicit instructions or delimiters provided to the agent to treat ingested test data as untrusted or to ignore embedded instructions within that data.
- Capability inventory: The skill workflow grants the agent capabilities to read/write project files, execute shell commands for reproduction, and run validation tools (SKILL.md).
- Sanitization: The instructions do not define any sanitization, escaping, or filtering steps for the data retrieved from failure logs or test files before the agent processes them.
Audit Metadata