The Agent Skills Directory

[SAFE]: The skill serves as a documentation-only resource, offering architectural guidance and Python code snippets. No executable malicious logic, obfuscation, or unauthorized data access patterns were identified within the files.
[SAFE]: The skill promotes security-positive coding practices, specifically recommending the use of Pydantic for strict validation of external data at system boundaries and the use of immutable records via frozen=True to prevent state corruption.
[SAFE]: Guidance for external service interaction includes industry-standard safety measures such as jittered exponential backoff and explicit, site-tuned timeouts to prevent resource exhaustion and ensure polite behavior towards target servers.
[INDIRECT_PROMPT_INJECTION]: The skill addresses the processing of untrusted data (e.g., API responses) but mitigates potential risks by instructing the agent to utilize Pydantic validation and schema enforcement. These recommendations provide clear boundaries for data ingestion, reducing the surface area for indirect injection attacks.

python-conventions