champion-move-outreach

Pass

Audited by Gen Agent Trust Hub on Mar 29, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill implements a professional B2B outreach workflow with multiple 'Human Checkpoints' that ensure all data processing and outreach drafts are reviewed by a human operator before execution.
  • [COMMAND_EXECUTION]: The skill utilizes standard tools for web searching and integrates with user-defined sales platforms (e.g., Apollo, Clearbit, Salesforce). These operations are restricted to the intended business purpose of the skill.
  • [DATA_EXFILTRATION]: The skill processes contact information, which is necessary for its function. Data exfiltration risks are mitigated by the structured review process and the use of well-known, legitimate service providers for outreach.
  • [PROMPT_INJECTION]: The skill handles untrusted input from external sources (people lists and web search results), which is a common surface for indirect prompt injection.
  • Ingestion points: Input lists from CRM systems, CSV files, and web search data about companies.
  • Boundary markers: None explicitly mentioned in the instructions.
  • Capability inventory: Capabilities include web-search, contact enrichment via external tools, and email/LinkedIn message drafting.
  • Sanitization: The instructions do not specify sanitization for ingested data.
  • Mitigation: The 'Human Checkpoint' after drafting (Step 4) serves as a primary control to ensure malicious instructions embedded in data are not executed or propagated.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 29, 2026, 05:17 PM