菜单
The single reference for writing Condor routines — fetching Hummingbot data, parallel calls, reports/charts, continuous loops, and candlestick charts. Routes to a companion file per topic.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
How total_amount_quote and initial_positions define a controller's isolated capital — the budget a single pmm_mister controller trades with, and how to seed it with base assets you already hold so each controller is independent from the wider portfolio.
Run the MM bot status report: running bots, open/hold-mode positions, closed position breakdown (TP/SL/Early/Hold), PnL, volume, and error summary.
Ready-to-deploy pmm_mister config profiles (aggressive / balanced / conservative) — full parameter coverage including spreads, effectivization times, tolerance, order types, skew, and global TP/SL.
End-to-end playbook for deploying a pmm_mister bot on a token — regime analysis, profile selection, config creation, and bot deployment.
On hyperliquid_perpetual, tokenized perp contracts (equities/pre-IPO names like SPCX-USD) are issued by a provider and require an issuer prefix in the trading pair (e.g. XYZ:SPCX-USD, not SPCX-USD).
Create and operate autonomous trading agents the minimal way — create the agent from just its role + purpose, prove it's alive by consulting it, then progressively improve it with routines and (optionally) a loop strategy.
| name | routine_cookbook |
| description | The single reference for writing Condor routines — fetching Hummingbot data, parallel calls, reports/charts, continuous loops, and candlestick charts. Routes to a companion file per topic. |
| when_to_use | Before implementing or debugging ANY routine. Read this first, then pull the specific companion file(s) for what your routine actually does (data, async, reports, continuous, charts). |
| source | agent:routine_builder |
The patterns for building routines, split into companion files so you load only what your task needs. Read this overview, then fetch the relevant file(s):
manage_skill(action="read_file", name="routine_cookbook", file="hummingbot_client.md")
| Your routine needs to… | Read |
|---|---|
| Fetch market data, candles, prices, order book, portfolio, executors | hummingbot_client.md |
| Make 4+ parallel API calls / bulk fetch many pairs / rate-limit | async_patterns.md |
| Produce a report — KPIs, tables, Plotly charts, rich inline output | report_builder.md |
| Run a continuous loop (monitor, tracker, alerts) until stopped | continuous.md |
| Render a candlestick chart, indicator overlay, or volume footprint | candles_chart.md |
Most routines need report_builder.md plus one or two others. A continuous
price monitor with a live dashboard, for example, reads hummingbot_client.md
continuous.md.report_builder.md.await; never time.sleep, only asyncio.sleep.None/missing keys, return error strings, never raise to the caller.manage_routines(action="run", ...)) and fix until the output is clean.These are starting patterns, not a bypass — running a routine still goes through the normal execution/confirmation controls.