taste-distiller
Pass
Audited by Gen Agent Trust Hub on Sep 5, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill's documentation includes installation commands referencing a GitHub repository (
github.com/leoluyi/skills). This repository is hosted on a well-known service and belongs to the skill's author, posing no security risk. - [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied text to identify stylistic patterns and anti-patterns during its interview process.
- Ingestion points: User examples of rejected AI outputs are collected in
SKILL.md(Stage B). - Boundary markers: The instructions do not use specific delimiters for the user-provided text examples.
- Capability inventory: The skill does not have access to external tools, network operations, or shell execution, neutralizing potential injection risks.
- Sanitization: The skill relies on model-level interpretation of user input without explicit sanitization logic.
Audit Metadata