The Agent Skills Directory

[PROMPT_INJECTION]: The skill contains only technical documentation and code snippets for stream parsing. There are no natural language instructions attempting to override agent behavior or bypass safety guardrails.
[DATA_EXFILTRATION]: No sensitive file paths, hardcoded credentials, or unauthorized network operations were found. The code demonstrates standard HTTP response handling using established Rust libraries.
[OBFUSCATION]: The content is presented in clear text with standard code formatting. No Base64, zero-width characters, or hidden data patterns were identified.
[REMOTE_CODE_EXECUTION]: The provided Rust code uses standard serialization and asynchronous stream processing patterns. It does not contain any functions like eval, exec, or shell command execution from untrusted input.
[EXTERNAL_DOWNLOADS]: No remote scripts or packages are downloaded or executed. The documentation refers to well-known Rust crates and LLM provider APIs (OpenAI, Anthropic, Google) in a purely informative context.

streaming-architecture