mcp2cli-runtime-api-tooling
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The instructions suggest using
uvxandnpxto download and run themcp2clipackage from theknowsuchagencyrepository. - [COMMAND_EXECUTION]: The tool's
stdiomode allows launching external processes via command strings, such asnpx @modelcontextprotocol/server-filesystem. - [DYNAMIC_EXECUTION]: A Python code example demonstrates the use of
subprocess.runto execute themcp2cliutility with dynamic arguments. - [INDIRECT_PROMPT_INJECTION]: The skill processes data from external OpenAPI and GraphQL APIs, as well as MCP servers, creating a potential vector for instructions hidden in that data to influence agent behavior.
- Ingestion points: Fetches remote API specifications and tool definitions from external servers.
- Boundary markers: No explicit delimiters or warnings to ignore embedded content are provided in the skill instructions.
- Capability inventory: The skill has access to the network and can execute local commands.
- Sanitization: No content filtering or validation of external API responses is mentioned.
Audit Metadata