writing-fragments
Pass
Audited by Gen Agent Trust Hub on Sep 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection because it reads back data from a file that may contain untrusted content.
- Ingestion points: The agent is instructed to re-read the markdown file from disk before every write operation to preserve user-made changes (
SKILL.md). - Boundary markers: The skill uses horizontal rules (
---) to separate different writing fragments. - Capability inventory: The skill's capabilities are strictly limited to reading from and writing to a single markdown file on the local filesystem. It lacks network access, subprocess execution, or other high-privilege tools.
- Sanitization: The instructions do not specify any sanitization or filtering for the content within the fragments, though the limited scope of the skill significantly mitigates potential risks.
Audit Metadata