iterate-until-verified
Pass
Audited by Gen Agent Trust Hub on Aug 25, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill operates by wrapping and iterating on user-defined tasks, which introduces a surface for indirect prompt injection where malicious instructions in the input could attempt to bypass the verification gates.
- Ingestion points: The core workflow in SKILL.md ingests the user's original task to form a 'task contract'.
- Boundary markers: The skill includes 'AUTHORITATIVE TASK' markers in its Compose mode to delimit user input.
- Capability inventory: The process utilizes agent capabilities such as code execution, research, and artifact generation based on the task.
- Sanitization: The instructions do not provide specific methods for sanitizing or filtering the content of the user-provided task.
Audit Metadata