The Agent Skills Directory

[SAFE]: The skill consists of high-quality documentation files (.md) covering various aspects of the Anthropic API, including Message Batches, Token Counting, and Admin APIs. No executable scripts, binaries, or active configuration files are present.
[EXTERNAL_DOWNLOADS]: All external references in the documentation point to official and trusted Anthropic domains (e.g., platform.claude.com, api.anthropic.com) or official SDK repositories on GitHub (e.g., github.com/anthropics/*). These are well-known, legitimate services.
[CREDENTIALS_UNSAFE]: Examples within the documentation use standard security practices, such as environment variable references ($ANTHROPIC_API_KEY) or clearly labeled placeholders (YOUR_API_KEY). No sensitive data or real secrets are hardcoded.
[PROMPT_INJECTION]: The skill includes behavioral instructions (e.g., requiring the use of the references/ directory via grep) to ensure accuracy. These instructions are standard for knowledge-retrieval tasks and do not attempt to bypass safety filters or override core agent behavior.

anthropic-api