skills/molechowski/agent-skills/res-x/Gen Agent Trust Hub

res-x

Pass

Audited by Gen Agent Trust Hub on Mar 14, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill downloads the 'uv' Python package manager from astral.sh, which is a well-known and trusted tool in the Python ecosystem. Documentation for installation is provided clearly in the README.
  • [CREDENTIALS_UNSAFE]: The skill implements a secure credential management workflow. It instructs users to store their xAI API key in a dedicated macOS keychain rather than hardcoding it or using environment variables. The Python script retrieves the key using the 'security' CLI tool.
  • [COMMAND_EXECUTION]: The script uses 'subprocess.run' to interact with the macOS keychain ('security' command). This is a legitimate use case for retrieving stored credentials and does not involve executing untrusted input.
  • [DATA_EXFILTRATION]: Network operations are restricted to the official xAI API endpoint (api.x.ai). No sensitive data (like the API key or local files) is sent to unauthorized external domains.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes external content (tweets and articles). While it has an ingestion surface, it acts as a pass-through tool, presenting the fetched content to the user without executing instructions contained within that data. The use of structured prompts for the xAI API further mitigates risks.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 14, 2026, 09:35 PM