writing-shape
Pass
Audited by Gen Agent Trust Hub on Sep 12, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted 'raw material' provided by the user to generate new documents. This creates a vulnerability surface where malicious instructions embedded in the raw text could potentially manipulate the agent's output or behavior during the article-shaping process.
- Ingestion points: The skill reads an input markdown file of 'raw material' (fragments, unstructured prose, or transcripts) to use as the source pile (SKILL.md).
- Boundary markers: While the skill differentiates between the 'pile' and the 'article,' it does not specify explicit delimiters or 'ignore' instructions to prevent the model from interpreting content within the pile as commands.
- Capability inventory: The skill has permission to read from disk and append content to new or existing article files based on user-defined paths (SKILL.md).
- Sanitization: There is no mention of sanitization, filtering, or validation of the input content before it is processed and written to the output file.
Audit Metadata