remove-ai-style
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted Markdown content provided by users, which creates an attack surface for indirect prompt injection. Maliciously crafted prose could contain embedded instructions intended to override the agent's behavior during the analysis or rewrite process.
- Ingestion points: Content from the user-provided
<article-path>or standard input is ingested directly into the agent's context and processed by theanalyze_ai_style.pyscript. - Boundary markers: The skill does not employ specific delimiters (e.g., XML tags or unique markers) or instructions to ignore potential commands hidden within the user-supplied text.
- Capability inventory: The skill performs file system operations (reading and direct editing) and executes a local Python analysis script that processes the full text of user documents.
- Sanitization: No sanitization, validation, or filtering of the input prose is performed to detect or neutralize instructions that might be interpreted as prompts.
Audit Metadata