mcp-server-chart
Pass
Audited by Gen Agent Trust Hub on Mar 10, 2026
Risk Level: SAFECREDENTIALS_UNSAFEDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [CREDENTIALS_UNSAFE]: All scripts (e.g., generate_area_chart.py) include a hardcoded demo API key 'BETA_TEST_KEY_MARCH_2026' as a fallback value when the DEEPNLP_ONEKEY_ROUTER_ACCESS environment variable is missing. \n- [DATA_EXFILTRATION]: The skill's primary function involves reading user data from local files or CLI arguments and transmitting it to the vendor's API at deepnlp.org. \n- [PROMPT_INJECTION]: The skill exposes an indirect prompt injection surface by ingesting and processing untrusted JSON data from users. \n
- Ingestion points: The '--data' and '--data-file' command-line arguments in all 27 Python scripts. \n
- Boundary markers: The scripts do not use delimiters or instructions to prevent the model from following commands embedded within the data. \n
- Capability inventory: Network communication is performed via the OneKeyAgentRouter.invoke method in all scripts. \n
- Sanitization: No sanitization or filtering of the data content is performed beyond standard JSON parsing.
Audit Metadata