craft-diorama-still-life
Pass
Audited by Gen Agent Trust Hub on Aug 27, 2026
Risk Level: SAFE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external article content to extract metaphors and titles. 1. Ingestion points: Data is read from user-provided local file paths (SKILL.md). 2. Boundary markers: Absent; the skill does not use specific delimiters to isolate source text from instructions. 3. Capability inventory: File system access for reading source documents and writing output Markdown files (SKILL.md). 4. Sanitization: Absent; extracted content is interpolated directly into prompt templates.
- [COMMAND_EXECUTION]: The agent is instructed to perform file system operations, specifically reading article content and writing generated prompt files. These are core functional requirements of the skill's described purpose.
Audit Metadata