writing-fragments
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructions mandate reading a markdown file from the local disk before every append operation to incorporate user edits. This behavior introduces a surface for indirect prompt injection, as any instructions contained within that file could be processed by the agent. \n
- Ingestion points: The markdown file specified by the user and read by the agent as per
SKILL.md. \n - Boundary markers: No specific delimiters or safety warnings are implemented to distinguish file content from system instructions. \n
- Capability inventory: The skill has the capability to read from and write to the local file system. \n
- Sanitization: No sanitization or filtering is applied to the content read from the file system.
Audit Metadata