The Agent Skills Directory

[COMMAND_EXECUTION]: The skill contains example code snippets for interacting with the Anthropic API and Redis. These are standard implementation patterns for the stated purpose of caching and do not involve arbitrary command execution or shell injection.- [DATA_EXFLLTRATION]: The skill demonstrates safe credential management by referencing process.env.REDIS_URL for database connections rather than hardcoding secrets. Network operations are directed towards well-known services (Anthropic API and Redis).- [CREDENTIALS_UNSAFE]: No hardcoded API keys, tokens, or passwords were found. All examples use environment variables or placeholders.- [REMOTE_CODE_EXECUTION]: The skill does not download or execute remote scripts. It uses standard Node.js imports for crypto and established packages for API interaction.- [PROMPT_INJECTION]: The skill contains technical instructions for LLM optimization and does not include markers or patterns intended to bypass AI safety filters or override system instructions.

prompt-caching