Skip to main content

morning-prep

Daily morning scan by EYE - load the saved morning layout, screenshot each watchlist symbol, and summarise overnight pre-market state from those charts. Use when the user says "good morning" or "morning prep". Do NOT use this when the user asks for a brief built from a named strategy's own signals or decision table - a strategy-specific brief skill owns that, and this one computes nothing from an indicator.

Ir a la instalación

Datos de origen

Repositorio
FerroxLabs/wayland
Última actividad en el origen
2 de septiembre de 2026 a las 16:20
Idioma detectado de SKILL.md
inglés
Estrellas
605
Forks
113

Opciones de instalación

De forma predeterminada está seleccionado el prompt que primero revisa el origen. Puedes cambiar a un comando directo o descargar una copia local.

Revisa los archivos de origen

Lee SKILL.md y los archivos complementarios que muestra SkillsMP antes de decidir si quieres instalarlo.

Mostrando SKILL.md

SKILL.md
Instrucciones de origen · Vista previa de solo lectura
name
morning-prep
description
Daily morning scan by EYE - load the saved morning layout, screenshot each watchlist symbol, and summarise overnight pre-market state from those charts. Use when the user says "good morning" or "morning prep". Do NOT use this when the user asks for a brief built from a named strategy's own signals or decision table - a strategy-specific brief skill owns that, and this one computes nothing from an indicator.
# Morning Prep Workflow You are running the user's daily pre-market routine on TradingView. ## Step 1: Pull the Watchlist 1. `watchlist_get` — read the active watchlist symbols. Membership comes from the symbols API, so it is complete even with the watchlist panel closed; prices are best effort and only attached when the panel is open, so check `quotes_available` before reading any price off it. 2. If empty, ask the user to import one via `watchlist_import` (they may have a JSON backup saved from a prior `watchlist_export`) Example watchlist: `["ES1!", "NQ1!", "CL1!", "GC1!", "DXY", "BTCUSD"]`. ## Step 2: Restore Morning Layout (if saved) 1. `state_list` — check for a snapshot **named** "morning" or similar 2. If found: `state_restore` with `name` set to that snapshot name — this brings back the user's preferred indicators, drawings, timeframe, and chart type. It is DESTRUCTIVE: it makes the chart match the snapshot, so anything not in the snapshot is removed. Take a `state_snapshot` of the current chart first if it might be wanted back. 3. If not found: proceed with current chart state and offer to `state_snapshot` with `name` "morning" at the end If `state_restore` fails with `CHART_LOADING`, wait ~2s and retry once. ## Step 3: Batch Screenshot the Watchlist Use `batch_run` with `action: "screenshot"` and the full symbol list. This captures each symbol without manual `chart_set_symbol` loops. ``` batch_run({ symbols: <watchlist>, action: "screenshot" }) ``` Screenshots save to `screenshots/` with timestamps — reference them in the summary. ## Step 4: Deep-Dive Top 5 Pick the 5 most-relevant symbols (user's main instruments, e.g., `ES1!`, `NQ1!`, `CL1!`, `GC1!`, `BTCUSD`). For each: 1. `chart_set_symbol` — switch the chart 2. `data_get_ohlcv` with `summary: true` — compact stats (overnight high/low, range, change%) 3. `quote_get` — the price held by the chart you just switched to. It is only as fresh as the chart's data connection and reports `success: true` on a stale bar, so check `tv_health_check` once before the loop and say so plainly if `datafeed.state` is `disconnected` rather than reading out prices that look current. `quote_batch` takes all five symbols in one call, does not touch the chart, and does not depend on that connection. Keep OHLCV summary-only to avoid context bloat. ## Step 5: Indicator Pulse For the top-5 only: 1. `data_get_study_values` — RSI, MACD, EMAs currently on chart 2. `data_get_pine_labels` with `study_filter` if the user runs a custom profiler (e.g., "Profiler", "Session Stats") If `data_get_indicator` is needed and returns `source: "dom_fallback"`, note that inputs were read from DOM (less reliable than API) — flag any suspicious readings. ## Step 6: Summarize to Chat Produce a compact briefing: - **Overnight leaders/laggards** — from batch screenshots + OHLCV change% - **Key levels in range** — pulled from `data_get_pine_lines` if a levels indicator is loaded - **Bias per instrument** — one line each (e.g., "ES1! — holding above PDH 5812, bullish above 5800") - **Watch items** — alerts firing near current price (use `alert_list` to cross-reference) ## Cleanup Offer to `state_snapshot` with `name` `morning-<date>` so today's layout is reproducible tomorrow. If the user added throwaway drawings while prepping, `draw_clear` them before snapshotting.
Ver en GitHub