The Agent Skills Directory

[Prompt Injection] (SAFE): No instructions attempting to bypass safety filters or override agent behavior were found.
[Data Exposure & Exfiltration] (SAFE): No hardcoded credentials, sensitive file access, or unauthorized network operations detected. Network requests in examples are directed to typical local endpoints like /api/chat.
[Obfuscation] (SAFE): No encoded or hidden content was found.
[Remote Code Execution] (SAFE): No patterns for downloading and executing untrusted remote scripts were identified.
[Indirect Prompt Injection] (LOW): The skill demonstrates processing untrusted streaming data from external APIs.
Ingestion points: response.body?.getReader() in SKILL.md and references/assistant-transport.md.
Boundary markers: Absent in the provided code snippets.
Capability inventory: The skill handles tool-call and text parts.
Sanitization: Not explicitly implemented in the provided code examples, which is standard for documentation focused on protocol implementation.

streaming