Agently-Skills

Use when developer-owned execution needs inspectable branching, concurrency, joins, waits, approval, retry, runtime streams, pause/resume, recovery, or mixed sync/async orchestration. Use agently-stage instead when a provider-owned sync wrapper only bridges an async SDK.

Use for Agently runtime extension surfaces: Actions and MCP, ExecutionResource lifecycle, TaskWorkspace file capabilities, RecordStore durability, auto_func and KeyWaiter helpers, FastAPIHelper transport, or optional agently-devtools observation and evaluation.

Use when shaping or collaboratively reviewing one Agently request family: model/settings setup, business Prompt contracts, structured output and validation, result or stream consumption, Session memory, embeddings, RecordStore retrieval, or retrieval-backed answers.

Use when migrating an existing LangChain, LangGraph, LlamaIndex, or CrewAI system into Agently and mapping each source behavior to its actual request, runtime, context/storage, or TriggerFlow owner.

Use when a model-powered product, assistant, automation, evaluator, or workflow request still needs the right Agently owner, execution shape, or project boundary chosen. Also use for explicit low-frequency TaskDAG or DynamicTask requests.