prospecting
Pass
Audited by Gen Agent Trust Hub on Sep 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes data from untrusted public sources (forums, reviews, social posts) to identify demand signals, particularly in the early-stage discovery branch.\n
- Ingestion points: Public discourse and repository issues are ingested via research tools such as last30days, social-fetch, and Firecrawl.\n
- Boundary markers: The instructions do not define specific delimiters or instructions to ignore embedded prompts within the external data when interpolated into the agent context.\n
- Capability inventory: The agent can write local CSV files and perform network operations via established sales intelligence APIs (Apollo, ZoomInfo, Truelist).\n
- Sanitization: The skill provides guidance to escape values when generating structured reports (JSON to HTML) to reduce injection risks.\n- [COMMAND_EXECUTION]: The skill uses a local CLI utility (github-prospects.js) to automate data extraction from GitHub. The agent executes this script using a static path and user-provided repository identifiers as part of the core prospecting functionality.
Audit Metadata