ai-loop
Pass
Audited by Gen Agent Trust Hub on Jul 9, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill utilizes user-provided requirements to drive a code generation loop, presenting a surface for indirect prompt injection. * Ingestion points: User interaction in the 'Spec' phase (SKILL.md). * Boundary markers: Requirements are formalized in a separate markdown file before implementation begins. * Capability inventory: Includes file writing and execution of verification commands. * Sanitization: Safety notes specify using sandboxed environments and requiring human approval for destructive or external actions.
- [COMMAND_EXECUTION]: The 'Review' phase involves running verification commands which are determined during the planning phase. The skill provides clear instructions to avoid unvalidated shell commands and to limit the execution scope to the implementation of the spec.
Audit Metadata