clay-performance-tuning
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external lead data (URLs, names, company info) from Clay tables. If this data contains malicious instructions, it could potentially influence the agent's behavior during enrichment or processing steps.
- [COMMAND_EXECUTION]: The skill requests broad Bash permissions via
curlto interact with APIs. While consistent with SaaS integration, it contributes to the vulnerability surface when processing untrusted input.
Audit Metadata