ulw-interview

Pass

Audited by Gen Agent Trust Hub on Sep 25, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted user input through an interview process and integrates these responses directly into local project artifacts.
  • Ingestion points: User input is gathered during the 6-round interview protocol described in SKILL.md.
  • Boundary markers: The instructions lack explicit delimiting or sanitization requirements for incorporating user-provided content into the clarified brief or existing artifacts.
  • Capability inventory: The skill has the capability to write and modify files as described in the 'Write the answer back' section of references/ambiguity-taxonomy.md.
  • Sanitization: No sanitization or validation of the user input is specified before it is integrated into persistent documentation.
  • [COMMAND_EXECUTION]: The skill utilizes platform-specific commands to track the status of the workflow.
  • Evidence: The SKILL.md file contains a shell command 'omh runtime record' used for execution tracking within the Oh My Hermes environment. This appears to be a status logging mechanism for the platform.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 25, 2026, 08:35 PM
Security Audit — agent-trust-hub — ulw-interview