states
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external content, such as component specifications or runnable previews, which serves as an ingestion surface for potential malicious instructions. * Ingestion points: External screens, components, or specs provided by the user for auditing (SKILL.md). * Boundary markers: The instructions lack explicit delimiters or warnings to prevent the agent from following instructions embedded within the processed data. * Capability inventory: The skill allows the agent to modify files ("implement the reachable states") and potentially interact with external systems to "replay" them (SKILL.md). * Sanitization: No sanitization or validation protocols are defined for the external inputs.
- [DYNAMIC_EXECUTION]: The skill instructs the agent to implement and "replay" states, which involves generating code and potentially executing it to verify behavior. * Evidence: The "Deliver" section states: "If fixes were requested, implement the reachable states and replay them." (SKILL.md).
Audit Metadata