enterprise-user-management-ai-analytics

Warn

Audited by Gen Agent Trust Hub on Jul 23, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill directs the user to clone a repository from an unverified third-party account ('Nareshkumar2583') on GitHub. This source is not associated with the skill author ('aradotso') or any recognized trusted organizations listed in the analysis framework.\n- [COMMAND_EXECUTION]: The instructions guide the user to execute shell commands such as 'npm install', 'pip install', and server start commands ('npm start', 'uvicorn') on the contents of the unverified third-party repository, which may lead to the execution of untrusted code.\n- [PROMPT_INJECTION]: The skill exhibits a vulnerability surface for indirect prompt injection. \n
  • Ingestion points: The 'classifyTicket' function in 'SKILL.md' ingests external data from support ticket subjects and descriptions provided by users.\n
  • Boundary markers: No delimiters or explicit instructions are provided to the agent or ML service to ignore potentially malicious commands embedded within the ticket data.\n
  • Capability inventory: The skill uses 'fetch' to send this untrusted data to an ML backend and receives automated results that could influence ticket classification or priority.\n
  • Sanitization: No input validation, escaping, or filtering of the ticket content is implemented before it is processed by the AI components.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Jul 23, 2026, 03:53 PM
Security Audit — agent-trust-hub — enterprise-user-management-ai-analytics