The Agent Skills Directory

[PROMPT_INJECTION]: The skill includes code patterns that illustrate potential vulnerability surfaces for indirect prompt injection.
Ingestion points: resources/implementation-playbook.md includes a get_user_data function that takes a user_id input.
Boundary markers: No delimiters or instructions are used to isolate the variable from the command context.
Capability inventory: The skill demonstrates simulated database access and query execution.
Sanitization: In the provided example, SQL queries are constructed using f-string interpolation, which is a significant security risk (SQL injection) if applied to production environments with untrusted input.
[EXTERNAL_DOWNLOADS]: The skill references network communication using established libraries.
Examples use aiohttp for making asynchronous HTTP requests to external domains.
The code includes references to well-known technology and testing resources such as httpbin.org and api.example.com.

async-python-patterns