panic-room-finder

Pass

Audited by Gen Agent Trust Hub on Mar 5, 2026

Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
  • [PROMPT_INJECTION]: Indirect Prompt Injection Risk. Ingestion points: The skill allows the use of the WebFetch tool for external data entry. Boundary markers: No delimiters or instructions are provided to distinguish external content from the skill's own logic. Capability inventory: The skill is permitted to use the Bash tool and file system editing commands (Read, Write, Edit). Sanitization: There is no instruction for sanitizing or escaping data retrieved from external sources.
  • [NO_CODE]: The skill is composed entirely of Markdown documentation and templates. No executable scripts in Python, Node.js, or other programming languages are included in the skill files.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 5, 2026, 02:23 PM