cold-email-sequence-generator-skill

Pass

Audited by Gen Agent Trust Hub on Oct 8, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to execute a local shell script scripts/variant-assigner.sh with a dynamic timestamp argument ($(date +%s)) to handle A/B testing variants.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external sources including HubSpot lists and CSV files to generate outbound email content.
  • Ingestion points: HubSpot list members, HubSpot contact search, and CSV file uploads (Stage 1).
  • Boundary markers: The workflow utilizes template brackets and a mandatory check_my_copy gate to review generated content.
  • Capability inventory: The skill has the ability to create Gmail drafts via mcp__claude_ai_Gmail__create_draft (Stage 6).
  • Sanitization: Employs a check_my_copy tool to validate every subject line and body variant against brand standards.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 8, 2026, 09:52 PM