writing-fragments
Pass
Audited by Gen Agent Trust Hub on Nov 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill reads external file content from the local disk back into the agent's context to maintain synchronization with user edits. This creates a potential attack surface where malicious instructions stored in the file could be executed by the agent.
- Ingestion points: Instructions in SKILL.md require the agent to re-read the file from disk before every write operation.
- Boundary markers: The skill lacks delimiters or instructions to ignore or escape instructions that might be contained within the fragments read from the file.
- Capability inventory: The agent can perform file system read and write operations on paths provided by the user, as specified in SKILL.md.
- Sanitization: No mechanisms are described to sanitize, validate, or filter the content retrieved from the file before it is processed.
Audit Metadata