The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill references documentation and SDK resources from modelcontextprotocol.io and the official modelcontextprotocol GitHub organization. These represent well-known and trusted sources for the Model Context Protocol framework.
[COMMAND_EXECUTION]: The evaluation harness script (scripts/evaluation.py) enables testing of local MCP servers using the stdio transport. This involves executing shell commands and arguments that are explicitly provided by the user during command-line invocation.
[PROMPT_INJECTION]: The skill exposes an indirect prompt injection surface through its evaluation harness, which processes external task descriptions and tool outputs.
Ingestion points: Data enters the system via evaluation XML files and results returned by the MCP tools under test.
Boundary markers: The system prompt in evaluation.py uses XML tags (<summary>, <feedback>, <response>) to structure and isolate agent outputs.
Capability inventory: The script can spawn subprocesses (for stdio servers) and perform network requests to the Anthropic API.
Sanitization: The script utilizes standard XML parsing libraries and regular expression extraction to manage data flow.

mcp-builder