openaccountants-us-ca
Pass
Audited by Gen Agent Trust Hub on Jun 20, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill invokes specific MCP tools including 'start', 'get_skill', and 'request_accountant_review' to interact with the vendor's tax logic engine. These are standard functional calls for an MCP-based agent skill.
- [DATA_EXFILTRATION]: When the user triggers a professional review, the skill transmits 'working papers' to the developer's verified domain (openaccountants.com). This behavior is consistent with the stated purpose of routing tax scenarios to human verifiers.
- [PROMPT_INJECTION]: The instructions include a mandate to prioritize retrieved tax rules over general training data. This is an accuracy-focused directive designed to ensure the agent uses current legislative rates and citations rather than potentially outdated internal knowledge.
Audit Metadata