build-prospect-list
Pass
Audited by Gen Agent Trust Hub on Oct 1, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes
clay whoamito verify the user's workspace and environment configuration during Step 0. This is a benign diagnostic operation within the vendor's ecosystem. - [INDIRECT_PROMPT_INJECTION]: The skill ingests external data from Clay's company and people search indexes, creating a surface for potential indirect prompt injection.
- Ingestion points: Results from
clay searchfor companies and people (SKILL.md). - Boundary markers: None explicitly defined in the prompt text to separate search data from agent instructions.
- Capability inventory:
clayCLI tools for searching and workspace verification, and managed enrichment functions (SKILL.md). - Sanitization: The skill enforces strict post-validation gates (Step 2, Step 5) that check returned data fields against specific ICP and persona criteria, which serves as a functional filter against malformed or unexpected data payloads.
Audit Metadata