openproject

Pass

Audited by Gen Agent Trust Hub on Mar 3, 2026

Risk Level: SAFE
Full Analysis
  • [DATA_EXFILTRATION]: The skill accesses a local .env file to retrieve OpenProject API credentials and stores a metadata cache in .openproject-config.yml. This is standard configuration and performance optimization behavior for this type of tool.\n- [EXTERNAL_DOWNLOADS]: The skill allows downloading file attachments from the user-configured OpenProject instance. The download locations are provided by the authenticated OpenProject API responses.\n- [SAFE]: The skill uses yaml.safe_load() for configuration parsing, preventing unsafe deserialization risks. Data presented to the agent is sanitized through field extraction in the op.py CLI tool.\n- [SAFE]: Indirect Prompt Injection surface analysis:\n
  • Ingestion points: Data from OpenProject API (e.g., Work Package subjects/descriptions) processed in op.py and sub-packages.\n
  • Boundary markers: Not present in the JSON output.\n
  • Capability inventory: File writing (download_attachment), network requests (httpx), and API-based data modification.\n
  • Sanitization: The skill extracts and structures specific fields before outputting JSON, which reduces the risk of the agent misinterpreting embedded instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 3, 2026, 07:02 AM