The Agent Skills Directory

[SAFE]: The skill is purely educational, providing documentation, templates, and examples for Python async/await patterns. It follows industry best practices for resource management, error handling, and concurrency control.
[REMOTE_CODE_EXECUTION]: No remote code execution or package installation commands were found. The skill only provides static code templates for common libraries like aiohttp and asyncpg.
[CREDENTIALS_UNSAFE]: While the templates contain placeholders like password="password", these are standard development defaults for local databases (e.g., PostgreSQL defaults) and do not represent a credential leak. The documentation correctly advises users to customize these parameters.
[DATA_EXFILTRATION]: There are no patterns suggesting data exfiltration. The HTTP requests demonstrated in templates/async-service-template.py and examples/fastapi-async-example.py are for legitimate API interaction (e.g., jsonplaceholder.typicode.com) used as placeholders for service integration.
[PROMPT_INJECTION]: The instructions do not contain any attempt to bypass agent safety filters or override system prompts. The allowed-tools field is correctly scoped to file reading operations necessary for a code-based skill.

python-async-patterns