The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill provides code patterns that ingest external data (URLs and endpoints) and perform network operations. While these are educational snippets, they represent a surface for indirect prompt injection if implemented in a system that processes untrusted data without validation.
Ingestion points: Functions such as fetch_data(url), resilient_api_call(endpoint), and robust_http_call(url) in SKILL.md accept external strings.
Boundary markers: None provided within the code snippets.
Capability inventory: Network access via the httpx library.
Sanitization: The provided patterns focus on resilience logic and do not include URL validation or sanitization logic.
[EXTERNAL_DOWNLOADS]: The skill references standard, well-known Python libraries such as tenacity, httpx, and structlog for implementation. These are legitimate dependencies for the stated purpose of the skill.

python-resilience