mix-engineer
Pass
Audited by Gen Agent Trust Hub on Feb 25, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection via its configuration loading mechanism.
- Ingestion points: The skill retrieves custom processing settings from an external file named "mix-presets.yaml" using the load_override tool.
- Boundary markers: The instructions do not specify any delimiters or instructions for the agent to disregard potential commands embedded within the YAML configuration.
- Capability inventory: The agent is equipped with several high-privilege tools, including "Bash", "Edit", and "Write", which could be exploited if the agent follows malicious instructions found in the ingested data.
- Sanitization: The skill documentation does not mention any form of schema validation or content sanitization for the loaded override files.
Audit Metadata