hybrid-executor
Pass
Audited by Gen Agent Trust Hub on Feb 24, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes external command-line interfaces for Claude and Gemini models via shell calls to perform specialized processing.
- [PROMPT_INJECTION]: The skill provides an indirect prompt injection surface by interpolating user input directly into sub-agent prompts. 1. Ingestion points: User input entering through the $ARGUMENTS variable in the shell commands defined in SKILL.md. 2. Boundary markers: None; the arguments are placed inside quotes within the prompt string without explicit delimiters. 3. Capability inventory: Shell access to execute claude and gemini binaries. 4. Sanitization: No sanitization, escaping, or validation is performed on the input before it is passed to the external models.
Audit Metadata