llm-text-hygiene
Pass
Audited by Gen Agent Trust Hub on Nov 13, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill's primary function is to ingest and process untrusted external data (finished texts, papers, and documentation) to audit them for AI artifacts. This ingestion creates a vulnerability where malicious instructions hidden within the analyzed documents could attempt to override the agent's behavior.
- Ingestion points: Deliverable files (e.g., .tex, .md, .txt) provided by the user for cleaning.
- Boundary markers: Absent. The instructions do not define delimiters or specific 'ignore embedded instructions' warnings for the agent when reading the provided content.
- Capability inventory: File system access to modify text content and the ability to execute CLI commands for text processing.
- Sanitization: Absent. There is no specified logic for sanitizing or escaping the content of the text files before they are processed by the agent.
- [EXTERNAL_DOWNLOADS]: The skill provides a link to an optional external tool,
pasta-press, hosted on the vendor's GitHub repository (https://github.com/ellmos-ai/pasta-press). This tool is intended for style refinement and is only used upon explicit user request. As a resource provided by the skill's author (ellmos-ai), it represents legitimate vendor-owned functionality.
Audit Metadata