startup-problem-ledger
Pass
Audited by Gen Agent Trust Hub on Oct 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied observations to create problem ledgers. This creates a surface for indirect prompt injection where malicious instructions in the input could attempt to manipulate agent output.\n
- Ingestion points: Raw user input is captured in the 'Workflow' and 'Problem ledger' sections of SKILL.md.\n
- Boundary markers: There are no explicit delimiters for user input within the processing logic, although HTML-escaping is mandated for artifact output.\n
- Capability inventory: The skill uses filesystem tools to write .html files.\n
- Sanitization: The instructions explicitly require HTML-escaping user-provided text to prevent script execution.\n- [DYNAMIC_EXECUTION]: The skill generates standalone HTML files that incorporate user-provided content.\n
- Evidence: The 'Optional HTML artifact' section instructs the agent to create and save .html files based on the problem ledger data. Generating renderable file formats from external input is a form of dynamic content generation.
Audit Metadata