playbook-new-in-role
Pass
Audited by Gen Agent Trust Hub on Aug 23, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [DATA_EXPOSURE_AND_EXFILTRATION]: The skill performs network operations to external services to fetch and process lead data. It communicates with the Prospeo API (api.prospeo.io) for lead searches and the OpenAI API (api.openai.com) for content generation, and it mentions the use of Apify for LinkedIn scraping.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external sources and incorporates it into LLM prompts. Ingestion points: Data returned from Prospeo and Apify APIs (e.g., job titles and names). Boundary markers: Uses a clear 'PER-ROW DATA' delimiter to separate system instructions from the lead data processed in the LLM prompt. Capability inventory: Performs network requests via HTTP and executes logic in Python workflow nodes. Sanitization: Includes specific instructions for the LLM to clean company names and follows a strict JSON output schema.
- [DYNAMIC_EXECUTION]: The workflow configuration includes Python code blocks for data processing in clay-workflow.md. These scripts (Nodes 2, 4, and 6) parse API responses, calculate job tenure, and perform data validation using standard Python logic.
Audit Metadata