writing-fragments
Pass
Audited by Gen Agent Trust Hub on Oct 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill maintains a workflow that involves reading and writing to a local markdown file based on user input, creating a surface for instruction injection.
- Ingestion points: The skill captures all user dialogue and re-reads the target markdown file from disk before each write operation to preserve external changes.
- Boundary markers: Fragments are separated by horizontal rules (---), which may not prevent the agent from interpreting instructions contained within the fragments during the re-reading phase.
- Capability inventory: The skill performs file read and write operations.
- Sanitization: No explicit sanitization or filtering is applied to the content before it is stored or re-processed.
- [NO_CODE]: The skill does not include any executable scripts or binary files, consisting entirely of markdown instructions and YAML configuration.
Audit Metadata