The Agent Skills Directory

[SAFE]: The skill serves as a high-quality documentation repository for the Anthropic Claude API and its associated SDKs.
[EXTERNAL_DOWNLOADS]: The documentation references external Model Context Protocol (MCP) servers and official Anthropic documentation hosted on GitHub and platform.claude.com. These are trusted sources and integral to the framework described.
[COMMAND_EXECUTION]: The Agent SDK examples demonstrate the use of a built-in Bash tool and mcpServers configuration for executing shell commands. These are documented features of the agentic framework intended for development tasks like browser automation and database access.
[PROMPT_INJECTION]: The skill includes instructions to prioritize specific model IDs and adaptive thinking modes. These are presented as best-practice defaults for the documented models and do not attempt to bypass safety guidelines or override system prompts in a malicious way.
[DATA_EXPOSURE]: Code examples correctly demonstrate using environment variables for API keys and database credentials, following secure development practices.

claude-api