fast-shot
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill establishes a workflow for processing user-supplied prompts as evidence, creating an attack surface for malicious instructions within that input.
- Ingestion points: The skill instructs the agent to 'Treat the prompt as evidence' and 'Preserve user wording, constraints, outputs' during the reasoning process (SKILL.md).
- Boundary markers: The skill encourages keeping 'USER_FACT', 'EXTERNAL_FACT', and other types distinct but lacks explicit delimiters or instructions to ignore embedded commands within those facts.
- Capability inventory: The 'Decide, act, and deliver' section directs the agent to 'perform the smallest authorized action' and 'run the smallest authorized empirical operation', implying the use of system tools (SKILL.md).
- Sanitization: The 'Integrity gate' provides a verification step stating that 'action never exceeds the user's scope or authority', which acts as a manual boundary, but no automated input sanitization is defined.
Audit Metadata