clay-toolkit

Pass

Audited by Gen Agent Trust Hub on Jul 18, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill includes a local Python script scripts/check-output.py designed to validate generated blueprints. The script reads file content based on command-line arguments.
  • [SAFE]: The skill creates a surface for indirect prompt injection by ingesting untrusted contact data (e.g., from CSVs or APIs) and processing it through AI-powered columns for personalization and email drafting. This is inherent to the skill's purpose for data enrichment.
  • Ingestion points: Contact data imported into Clay tables from Google Sheets, CSV, or external APIs as described in SKILL.md and references/gtm-table-blueprints.md.
  • Boundary markers: None identified in the provided templates to isolate external data from prompt instructions.
  • Capability inventory: The skill utilizes LLM-based columns for signal detection and email drafting, along with automated CRM push capabilities.
  • Sanitization: No specific sanitization or filtering logic is provided for the external contact fields before they are interpolated into AI prompts.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 18, 2026, 08:48 PM
Security Audit — agent-trust-hub — clay-toolkit