The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill provides patterns for making network requests (e.g., fetch_data(url: str)) that ingest external data, which is an inherent surface for indirect prompt injection if the agent processes the untrusted content returned by these URLs.
Ingestion points: The url and endpoint parameters in SKILL.md functions fetch_data, resilient_api_call, and http_request.
Boundary markers: The snippets focus on resilience logic and do not include prompt-level delimiters or instructions to ignore instructions embedded in the fetched data.
Capability inventory: Network access via httpx.get, httpx.post, and httpx.request.
Sanitization: No URL validation or content sanitization is demonstrated in the educational snippets.
[SAFE]: The skill promotes standard industry libraries (tenacity, httpx, structlog) and follows defensive coding practices for distributed systems (e.g., jitter, exponential backoff, and timeouts).
[SAFE]: No hardcoded credentials, sensitive file access, or obfuscation techniques were detected. The example URLs use the standard example.com domain.

python-resilience