The Agent Skills Directory

[SAFE]: The skill acts as an integration guide for the Anthropic Messages API, providing standard code snippets using the official anthropic SDK.\n- [SAFE]: Operational directives explicitly prohibit hardcoding API keys, instructing the agent to pull ANTHROPIC_API_KEY from the environment only.\n- [SAFE]: The instructions advocate for tool safety by requiring allowlists and human approval for sensitive capabilities, and the provided code examples demonstrate how to implement these restrictions.\n- [SAFE]: Content analysis across all files (SKILL.md, gemini.md, gpt.md) revealed no evidence of obfuscation, malicious command execution, or exfiltration attempts.

anthropic-api