taiwan-equity-research-coverage

Warn

Audited by Gen Agent Trust Hub on Mar 29, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructs the agent to clone an external repository (github.com/Timeverse/My-TW-Coverage) during the installation phase.
  • [REMOTE_CODE_EXECUTION]: The primary functionality relies on executing Python scripts provided within the cloned repository, such as update_financials.py, add_ticker.py, and audit_batch.py.
  • [COMMAND_EXECUTION]: The provided code examples demonstrate the use of subprocess.run and sys.executable to programmatically launch shell commands and script executions from within the Python environment.
  • [DATA_EXFILTRATION]: The skill uses the yfinance library to connect to external Yahoo Finance servers to retrieve market data. While this is expected behavior for the skill, it represents a network-enabled data flow.
  • [INDIRECT_PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection:
  • Ingestion points: Reads markdown ticker reports from the Pilot_Reports/ directory and data from enrichment.json.
  • Boundary markers: None identified; wikilink extraction via regex \[\[([^\]]+)\]\] directly parses content without isolation.
  • Capability inventory: Includes file writing (update_enrichment.py), network access (yfinance), and script execution (subprocess.run).
  • Sanitization: No sanitization or validation of the content within report files is mentioned before processing or potential AI consumption.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Mar 29, 2026, 07:29 AM