The Agent Skills Directory

[SAFE]: The skill acts as a capability reference and architectural guide for the Pipecat framework. All code snippets provided are standard examples of framework usage for building real-time AI agents.
[EXTERNAL_DOWNLOADS]: Fetches additional documentation from the official Pipecat project at https://docs.pipecat.ai/llms.txt. This is a well-known service domain for this framework and does not represent a security risk.
[PROMPT_INJECTION]: The skill describes an architecture that processes untrusted external data, creating a surface for indirect prompt injection.
Ingestion points: User audio via transport.input() and external data via weather_api.get_weather.
Boundary markers: Absent from the high-level code snippets.
Capability inventory: Orchestrates multimodal AI pipelines involving STT, TTS, LLMs, and external tool execution.
Sanitization: Not explicitly shown in the provided architectural examples.

daily