asap
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user input and explicitly instructs the agent to bypass standard deliberation steps ('Start immediately', 'Minimize planning', 'Keep narration minimal'). This creates a surface where embedded instructions in task data might be executed with less scrutiny.
- Ingestion points: SKILL.md (the
/asap <task>command and subsequent user requests). - Boundary markers: None defined to isolate the user task from the skill's operational instructions.
- Capability inventory: General tool execution, file access, and command-line operations as dictated by the surrounding environment and project patterns.
- Sanitization: No explicit sanitization or filtering of input data is mentioned, though the skill does state that safety controls must not be bypassed.
Audit Metadata