score-inbound-leads
Pass
Audited by Gen Agent Trust Hub on Oct 1, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute
clay whoamito verify the workspace and ensure the necessary CLI tools are present. This is a standard environment check for the Clay platform. - [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted external data (enriched lead lists) which serves as an injection surface.
- Ingestion points: Reads lead data (email, person, and company fields) from CSVs or tables.
- Boundary markers: None explicitly defined in the prompt templates.
- Capability inventory: Perform string normalization, arithmetic scoring, and environment verification via shell.
- Sanitization: The skill mitigates risks by requiring deterministic code for scoring rather than LLM reasoning, ensuring data fields cannot influence the logic through prompt injection. It also normalizes input fields (domains, headcount, titles) before processing.
Audit Metadata