ai-sdr-setup
Pass
Audited by Gen Agent Trust Hub on Jul 18, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [SAFE]: The skill instructions focus on establishing operational guardrails for AI agents, specifically recommending draft-only modes and human approval gates to prevent autonomous misbehavior.
- [PROMPT_INJECTION]: The skill defines a surface for indirect prompt injection by processing external buying signals (e.g., funding news, hiring data) to draft outreach messages.
- Ingestion points: External signal sources mentioned in SKILL.md (e.g., funding announcements, hiring roles).
- Boundary markers: The skill requires that "every personalized claim needs a source URL or verified data field" and utilizes explicit human review gates.
- Capability inventory: Message drafting and reply classification via the agent.
- Sanitization: Relies on mandatory human-in-the-loop approval for all messages before sending.
- [COMMAND_EXECUTION]: The provided scripts/check-output.py is a simple Python script used to validate the structure of generated markdown deliverables. It operates on local files provided via command line arguments and does not exhibit malicious patterns like network access, unauthorized file access, or privilege escalation.
Audit Metadata