diary
Pass
Audited by Gen Agent Trust Hub on Sep 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill requires the agent to record the "Verbatim prompt" and "Interpretation" of user requests into diary files within the project repository.
- Ingestion points: The skill processes all non-trivial user prompts as input for creating diary entries in SKILL.md.
- Boundary markers: There are no instructions to use specific delimiters or escaping mechanisms when writing untrusted user content into the markdown diary files.
- Capability inventory: The skill instructions involve file system operations (creating
docs/diary/and writing.mdfiles) and version control actions (committing code and diary together). - Sanitization: The skill explicitly forbids sanitization by requiring the prompt context to be recorded "verbatim" and "exactly as given," which ensures that any malicious injection payload in a prompt is preserved in the project documentation.
Audit Metadata