marketing-meta-ads-analyzer
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted external data in the form of Meta Ads Excel (.xlsx) reports, which could potentially contain malicious instructions intended to influence the agent's report generation or downstream tools.
- Ingestion points: SKILL.md and references/B.md detail the processing of user-uploaded Excel files containing ad names and performance metrics.
- Boundary markers: The skill requires a mandatory disclaimer in all output formats (references/I.2) to clarify that results are data-driven diagnostics and not direct commands.
- Capability inventory: The skill parses Excel data, generates complex document formats (HTML with React/Recharts, DOCX, PPTX), and chains results to other specialized agents like 'ai-slop-reviewer'.
- Sanitization: While the workflow involves post-processing by a 'humanize' agent, there is no explicit mention of sanitizing or escaping the raw cell content from the Excel files before interpolation into the final reports.
- [EXTERNAL_DOWNLOADS]: The skill provides an option to fetch real-time ad data directly from the official Meta Ads connector at mcp.facebook.com/ads using standard OAuth 2.0 authentication.
Audit Metadata