totp-mcp-server

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill explicitly exposes a tool (get_totp_code) that returns live TOTP codes into model-visible output, which requires the LLM to handle/output time-limited secret codes verbatim and therefore poses an exfiltration risk.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). Source is the user/agent prompt (outsider-authored free text) flowing into the MCP client’s tool-calling, where the model can be induced to call get_totp_code(account) after reading untrusted content; the runtime boundary explicitly warns that untrusted web/page/issue/dependency text in context can trigger tool calls.