find-roles
Pass
Audited by Gen Agent Trust Hub on Nov 7, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill fetches data from public Job Board APIs (Greenhouse, Lever, Ashby) and Hacker News (Algolia API). These are well-known services and the data fetched is limited to public job posting information.- [COMMAND_EXECUTION]: The skill includes a Python script (
scripts/ats_fetch.py) to automate ATS fetching. This script uses standard libraries (requests,yaml) and performs filtered reads of company-specific public job feeds based on a user-provided configuration file.- [INDIRECT_PROMPT_INJECTION]: The skill processes external data (job titles and descriptions) from various job boards. While this presents an ingestion surface for untrusted data, the skill implements specific filtering rules (keyword matching, liveness checks) and restricts the agent to a read-only table output, minimizing the risk of instruction execution from the processed content.- [SAFE]: The browser-based JobRight integration (resources/jobright.md) is explicitly restricted to read-only operations. It requires the user to handle authentication and forbids the agent from clicking 'Apply', entering credentials, or performing bulk scraping. These constraints prevent automated account actions or credential theft.
Audit Metadata