菜单
Maintain local operator dashboards in dashboard/ using Streamlit and Dash for candle inspection and indicator visualization. Use when improving manual observability, chart behavior, or live diagnostics from parquet-backed data.
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
Analyze and evolve the overall architecture of next-gen-algo-trading-bot across ingestion, aggregation, strategy execution, storage, APIs, and UI layers. Use when making cross-module changes, debugging data flow breaks between modules, planning new features, or assessing tradeoffs and risk before refactors.
Implement and maintain broker connectivity for Zerodha in brokers/ including auth, session retry behavior, order/position APIs, and historical data sync helpers. Use when touching login flow, order execution, broker resilience, or market data backfill behavior.
Operate and extend the Typer CLI entrypoints in cli/main.py for engine start/stop and timeframe worker execution. Use when adding commands, wiring new modules into runtime startup, or debugging startup/config/env failures.
Manage and validate runtime configuration schemas and YAML files under config/ plus root config.yaml for the trading engine. Use when adding config fields, fixing env interpolation, or resolving mismatches between configured and supported timeframes/indicators.
Implement and maintain the core orchestration and compute pipeline in core/ including base interfaces, engine actors, aggregation, timeframe generation, and indicator generation. Use when changing runtime data flow, worker behavior, or strategy execution mechanics.
Maintain live market data feed adapters in feeds/ with focus on Zerodha websocket integration and normalized tick payloads. Use when adding new feed providers, normalizing tick schema, or debugging live tick ingestion issues.
| name | ngat-dashboard |
| description | Maintain local operator dashboards in dashboard/ using Streamlit and Dash for candle inspection and indicator visualization. Use when improving manual observability, chart behavior, or live diagnostics from parquet-backed data. |
Support quick local introspection tools for candles and indicators separate from the production-facing Vue dashboard.
dashboard/visualize.py: Streamlit dashboard with candle/Keltner rendering.dashboard/dash_app.py: Dash dashboard with symbol/timeframe selection and overlays.day and higher frames are partially handled.