The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill references Model Context Protocol (MCP) documentation and SDKs from official sources including modelcontextprotocol.io and public GitHub repositories. These references are used to inform the agent about protocol specifications and best practices.
[COMMAND_EXECUTION]: The guide includes instructions for standard development and testing commands, such as using npm run build for compilation, npx @modelcontextprotocol/inspector for functional testing, and python -m py_compile for syntax validation.
[INDIRECT_PROMPT_INJECTION]: The skill suggests fetching documentation from external URLs to guide the implementation process. This creates a potential surface for indirect prompt injection if the external sources were compromised, though the targets are official project sites.
Ingestion points: Instructions in SKILL.md (Phase 1.2, 1.3) fetch protocol specifications and SDK READMEs.
Boundary markers: Absent. The skill does not explicitly instruct the agent to use delimiters or ignore embedded instructions when reading these docs.
Capability inventory: The skill describes local project build commands (npm, python) and local testing via the MCP Inspector.
Sanitization: None specified for the external documentation content.

mcp-builder