parallel
Pass
Audited by Gen Agent Trust Hub on Mar 14, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill workflow presents an indirect prompt injection surface when formatting instructions for sub-agents.
- Ingestion points: In
SKILL.md, Step 3 defines a prompt template that interpolates untrusted data sources such as{problem description},{list},{paste}(error messages), and{description}directly into the sub-agent's instructions. - Boundary markers: The template lacks clear delimiters or explicit instructions to treat the interpolated content as data rather than instructions, which could allow malicious commands embedded in logs or descriptions to influence sub-agent behavior.
- Capability inventory: The skill dispatches
general-purposesub-agents which typically have significant capabilities, including file system access and code execution. - Sanitization: There is no mention of sanitizing or validating external content before it is processed by the agent.
Audit Metadata