deep-analysis
Warn
Audited by Gen Agent Trust Hub on May 12, 2026
Risk Level: MEDIUMREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill dynamically loads and executes Python modules based on computed names in run_real_test.py and scripts/lib/pipeline/collect.py using importlib.import_module and import.
- [EXTERNAL_DOWNLOADS]: The skill contains logic in run.py and scripts/lib/playwright_fallback.py to automatically download and install Python packages from PyPI mirrors, the cloudflared binary from GitHub, and the Playwright Chromium browser.
- [COMMAND_EXECUTION]: The skill uses subprocess.run and subprocess.Popen in run.py to execute system commands for dependency management, installing system-level binaries (using sudo), and establishing network tunnels.
- [PROMPT_INJECTION]: The skill fetches content from financial news sites and social media platforms and incorporates this untrusted data into prompts used for agent-driven investment analysis, creating an attack surface for indirect prompt injection.
Audit Metadata