office-hours
Pass
Audited by Gen Agent Trust Hub on Oct 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided responses to its interrogation questions and includes them verbatim in its output brief. This creates a surface for indirect prompt injection if a user provides malicious instructions instead of legitimate answers. However, the skill lacks sensitive capabilities like file system modification, command execution, or network access, which limits any potential impact to the current conversation context.\n
- Ingestion points: User answers to the six-question interrogation within the skill's flow.\n
- Boundary markers: The output template uses markdown blockquotes to delimit user content in the final brief.\n
- Capability inventory: None. The skill generates text output and does not invoke external scripts or dangerous tools.\n
- Sanitization: No explicit sanitization of user input is performed.\n- [SAFE]: No malicious patterns, obfuscation, hardcoded credentials, or unauthorized network operations were identified in the skill instructions.
Audit Metadata