writing-fragments
Pass
Audited by Gen Agent Trust Hub on Aug 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill captures raw user input and writes it to a local markdown file, which is subsequently re-read into the agent's context to preserve manual edits. This creates a standard ingestion surface for untrusted data.
- Ingestion points: User conversation and the generated markdown file on disk (SKILL.md).
- Boundary markers: Fragments are separated by horizontal rules (
---) as described in the file format section of SKILL.md. - Capability inventory: The skill utilizes file read and write/append operations to manage the writing document (SKILL.md).
- Sanitization: No explicit sanitization or filtering is applied to the fragments before they are written or read back.
- [DATA_EXFILTRATION]: The skill reads and writes to a user-specified file path to store writing fragments. This behavior is the primary intended function of the skill and is triggered by direct user interaction.
Audit Metadata