dx-lesson
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided findings and appends them to a persistent file (foundation/lessons.md) intended for future consumption by the agent.\n
- Ingestion points: Data enters via the
[the finding]argument in SKILL.md or through direct user query if the argument is missing.\n - Boundary markers: The skill enforces a specific Markdown structure (## Title — Date) which provides basic structural delimitation for the appended content.\n
- Capability inventory: The skill has localized file-system write access (append-only) to the
foundation/directory.\n - Sanitization: The skill does not explicitly sanitize the input text, though its append-only nature and structured format mitigate some risk of file corruption.
Audit Metadata