dodo-mcp-usage
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The instructions direct the user to run
npx -y dodopayments-mcp@latestandnpx -y mcp-remote@latestto execute MCP servers. These packages originate from the vendor's official ecosystem or represent standard utility tools for the MCP protocol. - [DYNAMIC_EXECUTION]: The skill describes an
executetool that runs TypeScript code against the Dodo Payments SDK. According to the documentation, this execution is performed in an isolated sandbox to mitigate risks. - [INDIRECT_PROMPT_INJECTION]: The skill involves processing data from external sources, which constitutes a potential attack surface:
- Ingestion points: Documentation search results from
dodo-knowledgeand live account data (products, subscriptions, customers) from thedodopayments-apiserver in SKILL.md. - Boundary markers: The skill does not define specific technical delimiters for input data but provides high-level guidance for users to confirm operations.
- Capability inventory: The
executetool provides programmatic access to the Dodo Payments API, which is utilized for reading and writing financial data. - Sanitization: The instructions recommend returning trimmed summaries rather than full objects to limit data exposure in the agent context.
Audit Metadata