implement
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
- NO_CODE (SAFE): The file SKILL.md does not contain any executable scripts, shell commands, or binary data. It serves as a documentation-based instruction wrapper.
- PROMPT_INJECTION (LOW): The skill is vulnerable to Indirect Prompt Injection as it is designed to ingest and act upon user-defined plans. Evidence Chain: 1. Ingestion points: User-provided 'defined plan' described in SKILL.md metadata. 2. Boundary markers: Absent. 3. Capability inventory: The skill is intended to 'code' features based on input, implying high-capability tool use (file-write/exec) in the broader agent context. 4. Sanitization: Absent.
Audit Metadata