The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill provides installation instructions for two external NPM packages: xcodebuildmcp and @kimsungwhee/apple-docs-mcp. These resources are maintained by individual community members (cameroncooke and kimsungwhee) rather than recognized software organizations or the skill's vendor.- [REMOTE_CODE_EXECUTION]: The documentation includes configuration snippets that use npx -y to download and execute these MCP servers. While this is standard for MCP delivery, it involves the execution of remote scripts from unverified third-party sources on the host system.- [INDIRECT_PROMPT_INJECTION]: The skill creates a vulnerability surface by instructing the agent to ingest and process large amounts of external content, including framework documentation and WWDC session transcripts (2014-2025).
Ingestion points: Data is pulled through the apple-docs MCP server tools (e.g., Search Documentation, Search WWDC Sessions, and Get Sample Code).
Boundary markers: The instructions do not specify any delimiters or boundary markers to help the agent distinguish between its system instructions and the untrusted documentation content it retrieves.
Capability inventory: The agent is granted significant system capabilities through the XcodeBuildMCP server, including discovering projects, performing build operations, and cleaning directories, which could be misused if influenced by injected data.
Sanitization: There is no evidence of sanitization, filtering, or validation performed on the retrieved documentation data before it is processed by the agent.

mcp-tools