wind-full-funnel
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFENO_CODEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill consists of Markdown files, a JSON metadata file, and a text identifier. It does not contain any executable scripts, binaries, or active components.
- [NO_CODE]: No programming code (Python, JavaScript, Shell, etc.) is included. The skill operates by providing the agent with natural language procedures and templates for marketing analysis.
- [INDIRECT_PROMPT_INJECTION]: The skill establishes a data ingestion surface by instructing the agent to process commercial briefs, budget data, and advertising performance signals (Step 1 and Step 3 in SKILL.md). However, the risk is addressed through explicit procedural instructions:
- Ingestion points: Commercial briefs and available account evidence.
- Boundary markers: Use of a structured output template (templates/output-template.md) and explicit citation rules (references/sources.md).
- Capability inventory: The skill is restricted to strategy generation; it contains an explicit warning that a strategy request does not authorize account changes or spending.
- Sanitization: Instructions emphasize recording assumptions and marking unknown values instead of fabricating data.
Audit Metadata