clay-policy-guardrails

Pass

Audited by Gen Agent Trust Hub on Mar 24, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill provides comprehensive logic for security and cost controls, including credit spending limits, PII data filtering, and data retention policies. It promotes security best practices by suggesting pre-validation and access controls for data enrichment pipelines.- [PROMPT_INJECTION]: The skill defines functions that process untrusted external data (enrichment rows) for validation, which constitutes an indirect prompt injection surface. While the current logic is defensive, processing attacker-controlled data always carries a risk if that data is subsequently interpolated into LLM prompts without sanitization.\n
  • Ingestion points: validate_rows_for_enrichment function in SKILL.md.\n
  • Boundary markers: Absent.\n
  • Capability inventory: The skill utilizes file system access (Read, Write, Edit) and shell execution capabilities (Bash).\n
  • Sanitization: The provided code implements basic regex validation for emails and string length checks for domains, which helps mitigate accidental data issues but may not prevent sophisticated injection attacks.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 24, 2026, 04:49 PM