gahan-b2b-creator-blueprint
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill's documentation and derivation metadata reference external URLs from LinkedIn and Favikon. These references are used neutrally to establish the provenance of the campaign methodology and provide research context from well-known professional services.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process external inputs such as customer research, product evidence, and creator analytics. It implements several safeguards to mitigate risks from untrusted data:
- Ingestion points: Processes user-supplied research and analytics data within the planning workflow.
- Boundary markers: Utilizes a strict output template (templates/output-template.md) and a comprehensive review checklist (checklists/review-checklist.md) to separate evidence from assumptions.
- Capability inventory: The skill is restricted to planning and drafting; it lacks capabilities for direct network transmission or file system modification outside of its own directory.
- Sanitization: Explicitly instructs the agent to prepare materials as drafts and forbids sending outreach, publishing content, or authorizing spend without separate, manual authorization.
Audit Metadata