The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill fetches official framework documentation and specification files from the Model Context Protocol website and GitHub repositories to provide the agent with technical context.
[REMOTE_CODE_EXECUTION]: Provides instructions for testing implementations using npx @modelcontextprotocol/inspector. This tool is the official utility for validating MCP servers and is treated as a standard developer dependency.
[COMMAND_EXECUTION]: Includes standard terminal commands for building and checking code, such as npm run build and python -m py_compile, which are typical for software development workflows.
[INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection due to its reliance on fetching external documentation via WebFetch. However, because these sources are the official project repositories, the potential for malicious content is considered minimal.
Ingestion points: SKILL.md (Phase 1.2, 1.3, and Documentation Library) defines URLs for fetching external markdown files.
Boundary markers: The skill does not define specific delimiters for external content, but the context is strictly technical reference for code generation.
Capability inventory: The skill allows for writing code to the file system and executing local build and test commands.
Sanitization: Content from external sources is processed without specific filtering, which is consistent with handling technical documentation.

mcp-builder