Menu
Async/await rules for echo-sdk — never block the loop, cancellation safety, asyncio.gather over sequential awaits. Use any time you write or modify async code in src/echo/.
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Based on SOC occupation classification
Router for any non-trivial echo-sdk task. Maps user intent to the right module skill and diagram. Invoke FIRST for any change to src/echo/ before reaching for a specific module skill.
Recipe for adding a new provider (LLM, transcriber, eval, prompt). Use whenever adding Cohere, Mistral, Deepgram, Phoenix, Promptlayer, etc. — anything that plugs into an existing get_*() factory.
End-to-end recipe for defining a Skill bundle and wiring it into a GenericAgent with the right activation mode and tool registry. Use when extending agent capability via swappable skills rather than baking tools in.
Building agents with BaseAgent / GenericAgent, including skill registration, activation modes (llm vs manual), and system-prompt composition. Use when editing src/echo/agents/ or building a new agent subclass.
Audio transcription provider abstraction (Gemini, EkaCare). Use when transcribing audio, adding a new transcriber, or shaping AudioInput.
Postgres support — asyncpg client, schema-aware binder, PgQueryTool surface. Use when querying postgres from a tool, modifying the binder, or adding a new DB engine.
| name | python-async-discipline |
| description | Async/await rules for echo-sdk — never block the loop, cancellation safety, asyncio.gather over sequential awaits. Use any time you write or modify async code in src/echo/. |
Echo is async-first. Sync code in an async path is a bug — it stalls every other in-flight request on the same loop.
httpx.AsyncClient, asyncpg, aiofiles). If you must run sync code, wrap with await asyncio.to_thread(fn, *args).time.sleep is forbidden in async paths. Use await asyncio.sleep(...).asyncio.gather — don't sequentially await two independent calls.async with for httpx.AsyncClient, asyncpg.create_pool, MCPConnectionManager. If you can't use async with, put await x.close() in finally.asyncio.CancelledError — re-raise after cleanup. Wrap teardown in try/finally, not try/except.asyncio.run inside library code. It belongs in entrypoints (examples/, tests/, host apps).aclose() if they stop iterating early. Prefer async for with async with of a contextmanager that owns the generator.asyncio.to_thread (preferred) or loop.run_in_executor — don't spin up bare threading.Thread.requests.get(...) in an async function → use httpx.AsyncClient.time.sleep → await asyncio.sleep.for x in items: result = await foo(x) where each call is independent → asyncio.gather(*(foo(x) for x in items)).try/except Exception: pass around an async call → swallowing cancellation.asyncio.run(...) inside src/echo/ → only in entrypoints.[[python-memory-safety]] (resource leaks), [[echo-sdk-tools]] (tool.run is async).