TradeOS analyze

TradeOS exposes 5 MCP tools on https://ai.tradeos.xyz. This skill maps user intent to the correct tool, argument shape, and analysis playbook.

Contents

1. Prerequisites

1. TradeOS MCP is connected — run mcp_health first after OAuth.
2. Never pass userId in tool arguments (OAuth injects it).
3. For stdio bridge only: TRADEOS_ACCESS_TOKEN must be set.

2. Tool picker

Default call chain: search_tickers → customize-agent (optional) → technical_analysis

3. Tool usage

Each subsection below is a how-to for one tool (mcp_health omitted — call with no args after OAuth).

3.1 search_tickers

Purpose: Symbol lookup only. Read-only — no agents, LLM, or charts.

When to use

• User gives a company name, alias, or partial ticker.
• Disambiguate before customize-agent or technical_analysis.
• Resolve each leg of a spread (assetA and assetB separately).

When NOT to use

• Chart TA → technical_analysis
• Agent CRUD → customize-agent
• News / macro → bloomberg-oracle-terminal

How to call

{ "q": "apple", "limit": 5, "exchange": "NASDAQ" }

Spread / ratio: search each leg, then join as "<tickerA> / <tickerB>" (see §4.6).

Response: ok, query, count, data[] (ticker, name, exchange, score, isWatch, …).

3.2 customize-agent

Purpose: My Agent CRUD + read last stored analysis from DB. Does not run a new LLM — use technical_analysis for live analysis.

When to use

When NOT to use

• Fresh chart write-up → technical_analysis
• Symbol lookup → search_tickers

How to call — list agents

{ "action": "list" }

How to call — create agent

{
  "action": "create",
  "create": {
    "ticker": "NASDAQ:AAPL",
    "interval": "1D",
    "extraWindow": ["1W", "4W"],
    "strategy": "Trend-Analysis",
    "indicators": [
      "Moving Average",
      "Bollinger Bands",
      "Relative Strength Index"
    ],
    "prompt": "Daily trend and key levels for AAPL.",
    "trigger": { "condition": [] }
  }
}

How to call — update / delete / get

{
  "action": "update",
  "update": { "agentId": "<id>", "strategy": "Smart-Money-Concept" }
}

{ "action": "delete", "delete": { "agentId": "<id>" } }

{ "action": "get", "get": { "agentId": "<id>" } }

How to call — read stored report (not fresh analysis)

{
  "action": "latest_analysis",
  "latest_analysis": { "agentId": "<id>" }
}

After create: run technical_analysis with action=agent and customizeAgentId from create/list response.

Rules

• Do not pass title on create (server-generated).
• update requires agentId + at least one mutable field; if indicators is sent, must be non-empty.
• Spread ticker: one string "assetA / assetB" (see §4.6). Pro users keep full spread on create/update.

3.3 technical_analysis

Purpose: Live Ticker-Analyze — server captures multi-timeframe chart screenshots for the LLM. Returns streamPreview (Markdown) + screenshotHints[] (text only).

When to use: Chart TA, levels, strategy framing, trade questions.

When NOT to use: Agent CRUD / stored report → customize-agent · symbol search → search_tickers · news → bloomberg-oracle-terminal.

Two modes

Time windows — common usage

1. Set `mainWindow` to the horizon the user cares about (usually 1D for equities).
2. Add `extraWindows` top-down for context — e.g. ["4W", "1W", "1D", "60"] (max 8 on ad_hoc).
3. Match strategy to primary window — e.g. Scalping-EMA on 5/15; Trend-Following on 1D/1W; Smart-Money-Concept on 60/240/1D.

Valid interval enum values: see §4.1.

#### 3.3.1 Single asset

One symbol (not a spread). Example: NASDAQ:AAPL, X:BTCUSD.

Path A — ad_hoc (one-off, no saved agent)

{
  "action": "ad_hoc",
  "ad_hoc": {
    "ticker": "NASDAQ:AAPL",
    "mainWindow": "1D",
    "extraWindows": ["1W", "4W"],
    "strategy": "Trend-Analysis",
    "indicators": [
      "Moving Average",
      "Moving Average Convergence Divergence",
      "Relative Strength Index",
      "Bollinger Bands"
    ],
    "userQuestion": "Is the daily structure still bullish? List key support/resistance, momentum confirmation, and volatility regime."
  }
}

Path B — agent (reusable My Agent)

{
  "action": "agent",
  "agent": {
    "customizeAgentId": "<from create or list>",
    "userQuestion": "Trend and key levels on the daily chart?"
  }
}

Suggested `userQuestion` checklist (single asset)

1. Primary trend on mainWindow (bullish / bearish / range)
2. Key support/resistance and invalidation level
3. Momentum confirmation or divergence
4. Volatility regime (compression vs expansion)
5. _(Optional)_ Higher-TF alignment from extraWindows

#### 3.3.2 Multi asset (spread vs benchmark)

Compare many symbols against one benchmark using ratio charts. One ticker per pair: "<symbol> / <benchmark>".

Common benchmarks

Workflow

Example payload (one pair)

{
  "action": "ad_hoc",
  "ad_hoc": {
    "ticker": "LLY / SPY",
    "mainWindow": "1D",
    "extraWindows": ["1W", "4W"],
    "strategy": "Trend-Analysis",
    "indicators": [
      "Moving Average",
      "Relative Strength Index",
      "Moving Average Convergence Divergence",
      "Bollinger Bands",
      "Average True Range"
    ],
    "userQuestion": "Analyze this spread for trend, momentum, and volatility. Score each 0–100. Summarize bias: long spread, short spread, or neutral."
  }
}

Scoring dimensions (state explicitly in userQuestion)

Example user prompts

1. _“Analyze NVDA/QQQ spread for trend, momentum, volatility.”_
2. _“Analyze LLY, GOOG, AMZN, META, MSFT, GLD, QQQ against SPY (LLY / SPY, …), rank 0–100.”_

Prefer one `ad_hoc` call per pair (parallel if host allows). Each call gets its own screenshots and cleaner streamPreview.

Multi asset via saved agents: create one My Agent per spread pair (customize-agent create with "LLY / SPY"), then technical_analysis agent for each — same scoring logic in userQuestion.

3.4 bloomberg-oracle-terminal

Purpose: Grounded search for macro / news / policy narrative. Returns Markdown summary — not chart TA.

When to use

• Fresh headlines, sector or event context (“why did oil move”, Fed news)
• Catalyst context before or after TA (not instead of chart analysis)

When NOT to use

• Per-symbol chart analysis → technical_analysis
• Ticker lookup → search_tickers
• My Agent CRUD → customize-agent

How to call

{
  "query": "Why did NVDA drop this week? Include earnings, AI sector news, and Fed context."
}

Response: ok, data.summary, data.query. Balance checked per account. No screenshots; host may timeout on slow searches.

Combined workflow (news + chart):

1. bloomberg-oracle-terminal — “why”
2. technical_analysis — “where on the chart”

4. Parameters and enums

All enum values must match exactly (case-sensitive). Invalid values → invalid_input.

4.1 Intervals (interval, mainWindow, extraWindows, extraWindow)

1 | 5 | 15 | 30 | 60 | 240 | 1D | 1W | 4W

Limits: extraWindows max 8 on ad_hoc; extraWindow optional on agent create.

4.2 customize-agent actions

create | update | delete | get | list | latest_analysis

4.3 technical_analysis actions

agent | ad_hoc

Optional (both modes): clientLang, selectedVisibilityType

4.4 Strategy enum

Use with customize-agent create/update and technical_analysis ad_hoc. 28 values — spelling and hyphenation must match exactly.

1. `Quantum-Athena` · Systematic

• User wants a bundled multi-factor scan without picking a manual setup.
• Screening many symbols with TradeOS systematic rules.
• Portfolio-level ranking where a pre-built logic stack is preferred.

2. `Chiroptera` · Systematic

• Recurring chart pattern workflows (flags, bases, recurring setups).
• User asks for pattern-based analysis with a fixed TradeOS template.
• Repeatable agent config across similar-looking charts.

3. `Quantum-Valkyrie` · Systematic

• Strong, fast momentum markets (crypto rips, parabolic leaders).
• User accepts higher drawdown for aggressive trend capture.
• Breakout continuation after high relative volume.

4. `Ultimate-Breakout-System` · Volatility / breakout

• Range compression followed by expansion across a watchlist.
• User asks “what is breaking out today” or systematic breakout scan.
• Volatility expansion after low-ATR consolidation.

5. `Technical-SEPA` · Trend / momentum

• US growth stocks in stage-2 uptrend (Minervini / SEPA style).
• Tight base near highs before potential continuation.
• User focuses on leadership names with clean trend structure.

6. `Trend-Analysis` · Trend / momentum

• Default choice for general TA questions.
• Single-symbol trend, levels, and multi-TF structure.
• Spread vs benchmark (AAPL / SPY) for relative strength.
• Multi-symbol ranking vs one benchmark (trend / momentum / vol scoring).

7. `Smart-Money-Concept` · SMC / structure

• User mentions order blocks, liquidity, ICT, BOS, CHoCH, FVG.
• Institutional narrative: where stops sit, where price may react.
• 60 / 240 / 1D structure with precise level framing.

8. `Liquidity-Sweep` · SMC / structure

• Sharp wick through prior high/low then reversal.
• “Stop hunt” or sweep-then-reverse explanation.
• Session open / news wick that takes liquidity and fades.

9. `Pullback-Retracement` · Pullback / range

• Established uptrend or downtrend; user wants dip-buy or rally-fade entry.
• Pullback to MA, Fib zone, or prior breakout level.
• Trend continuation after first retrace.

10. `Scalping-EMA` · Systematic

• Intraday trades on 1 / 5 / 15 intervals.
• Quick in-and-out around EMA dynamic support/resistance.
• User asks for scalp entries, tight stops, session trades.

11. `Volatility-Breakout` · Volatility / breakout

• Price exits a range or band with volume/vol expansion.
• Post-earnings or post-news range break.
• User asks whether a breakout is valid vs false break.

12. `Breakout-Retest` · Pullback / range

• Level broken, user waits for retest and hold before entry.
• Confirmation-style breakout (not first touch only).
• Failed retest = invalidation scenario.

13. `Squeeze-Momentum` · Volatility / breakout

• Bollinger/Keltner squeeze visible; pre-breakout coil.
• User asks “is volatility compressing” or “squeeze about to fire”.
• Pair with Squeeze Momentum and Bollinger Bands indicators.

14. `Mean-Reversion` · Pullback / range

• Range-bound market; fade moves to range mid or opposite band.
• Overextended move vs mean (often with Z-Score / RSI).
• Not ideal when ADX high and trend strong.

15. `Momentum-Swing` · Trend / momentum

• Hold horizon roughly 2–10 days on 1D / 1W.
• Trend intact plus momentum confirmation (MACD/RSI aligned).
• Swing entry after pullback in directional market.

16. `Trend-Following` · Trend / momentum

• Clear HH/HL or LH/LL structure; user wants to ride the trend.
• Trailing stop / stay-with-trend bias.
• Avoid in choppy sideways chop.

17. `Trend-Reversal` · Trend / momentum

• Climactic volume, exhaustion, potential major top or bottom.
• User asks “is the trend ending” or reversal candidates.
• Higher false signal rate — needs clear exhaustion cues.

18. `Divergence-Play` · Trend / momentum

• RSI or MACD divergence vs price (higher high / lower high mismatch).
• Early warning of weakening trend before structure breaks.
• User asks specifically about bullish/bearish divergence.

19. `Continuation-Pattern` · Trend / momentum

• Flag, pennant, wedge, or triangle after impulse leg.
• User asks “is this a bull flag” or pause-before-continue setup.
• Entry on pattern break in direction of prior trend.

20. `Range-Bound` · Pullback / range

• Horizontal channel with defined support and resistance.
• User trades bounces inside range until breakout.
• Explicit range high/low and midline context.

21. `Multi-TF-BOS-Alignment` · SMC / structure

• Break of structure on 2+ timeframes in same direction.
• High-confluence continuation after aligned BOS.
• User wants top-down TF agreement before taking risk.

22. `SR-Failure-Trap` · SMC / structure

• False breakout above resistance or below support.
• Fade failed breaks in ranging conditions.
• Trap / stop-run at obvious level then rejection.

23. `SuperTrend-Ladder-Pullback` · Systematic

• SuperTrend defines trend direction; entries on pullbacks to ladder levels.
• Mechanical trend-follow with defined add/trim zones.
• User wants rule-based SuperTrend workflow.

24. `Tri-MA-Slope-Compression-Break` · Systematic

• Three moving averages compress then slope expands together.
• Early signal of volatility regime shift.
• Pre-breakout MA ribbon compression on 1D or 240.

25. `MACD-Zero-Line-Recycle-Gate` · Systematic

• MACD above/below zero as momentum regime filter.
• Only long when MACD holds above zero (or inverse for shorts).
• User asks about MACD zero-line recycle or momentum gate.

26. `RSI-Range-Shift-Continuation` · Systematic

• RSI shifts into a new bull/bear range (e.g. 40–80 bull zone).
• Momentum persistence after regime shift, not single OB/OS tick.
• Continuation trades in strong momentum names.

27. `Squeeze-Fakeout-Reversal` · Volatility / breakout

• Squeeze fires but price fails to follow through.
• Counter-trend or fade after false expansion.
• User suspects bull/bear trap after squeeze break.

28. `Fractal-Sync-Confluence` · Trend / momentum

• 4W / 1W / 1D (or similar stack) align same bias.
• Fewer trades, higher confluence requirement.
• User explicitly wants multi-TF agreement before acting.

Quick picks

4.5 Indicator enum

English full names only — no abbreviations (e.g. not RSI, use Relative Strength Index). My Agent: 1–20; ad_hoc: up to 50. 26 values — must match exactly.

1. `Moving Average` · Trend

• Primary trend direction (price above/below MA).
• Dynamic support/resistance on pullbacks.
• Spread vs benchmark: ratio above/below MA for relative trend.
• Works on all horizons; 1D/1W most common for equities.

2. `Exponential Moving Average` · Trend

• Faster reaction than simple MA; intraday and crypto bias.
• Short swing entries when price holds above rising EMA.
• Scalp strategies (Scalping-EMA) and reactive trend filter.

3. `Average Directional Index` · Trend

• Measures trend strength, not direction.
• Filter: prefer trades when ADX > 25; skip low ADX chop.
• Confirm whether Trend-Following vs Range-Bound fits the chart.

4. `Bollinger Bands` · Volatility

• Band width for squeeze vs expansion.
• Mean reversion at outer band in ranges; breakout on band walk in trends.
• Volatility scoring on spread charts (benchmark analysis).

5. `Volume` · Volume

• Confirm breakout validity (rising volume on break).
• Spot climactic selling/buying on reversal candidates.
• Basic participation check on any strategy.

6. `Price Volume Trend` · Volume

• Cumulative volume-direction flow.
• Divergence: price new high but PVT flat/lower (weak rally).
• Trend health check alongside price structure.

7. `Relative Volume` · Volume

• Today vs average volume — unusual activity filter.
• Earnings, FDA, macro news days; opening drive strength.
• Intraday breakout confirmation.

8. `Moving Average Convergence Divergence` · Momentum

• Momentum shifts, signal-line crosses, histogram expansion.
• Spread momentum: ratio MACD vs benchmark direction.
• Pair with Trend-Analysis, Momentum-Swing, Divergence-Play.

9. `Ichimoku Cloud` · Trend

• Cloud support/resistance and trend bias (above/below cloud).
• 1D / 1W position trades and swing holds.
• Multi-component trend filter (tenkan, kijun, cloud).

10. `KDJ Oscillator` · Momentum

• Fast stochastic-style signals on short TFs (5/15/60).
• Reactive overbought/oversold in volatile sessions.
• Asian session equities and quick momentum turns.

11. `Relative Strength Index` · Momentum

• Overbought (>70) / oversold (<30) context.
• Bullish/bearish divergence vs price.
• Mean-reversion fades or Divergence-Play / RSI-Range-Shift-Continuation.

12. `Squeeze Momentum` · Momentum

• Momentum build after volatility compression.
• Pair with Bollinger Bands and Squeeze-Momentum strategy.
• Pre-breakout color/signal shift before range expansion.

13. `Support and Resistance` · Levels

• Horizontal key levels on any symbol or spread ratio.
• Universal add-on for entries, stops, targets.
• Required for Breakout-Retest, Range-Bound, SR-Failure-Trap.

14. `High Low Markers` · SMC

• Mark swing highs and lows for structure mapping.
• Liquidity context for Liquidity-Sweep and SMC workflows.
• Define dealing range boundaries.

15. `Fair Value Gap` · SMC

• Imbalance / gap zones for SMC entries and fill targets.
• Smart-Money-Concept and post-impulse retrace levels.
• Mitigation vs rejection at FVG.

16. `Fibonacci Retracement` · Levels

• Pullback depth (38.2%, 50%, 61.8%) in trends.
• Pullback-Retracement entry zone selection.
• Confluence with MA or S/R for dip buys.

17. `Z-Score` · Volatility

• Statistical distance from mean on spread or single symbol.
• Extreme Z → mean-reversion candidate.
• Pair/spread ranking for stretched relative value.

18. `Volume Profile Visible Range` · Volume

• High-volume nodes (HVN) and low-volume gaps (LVN).
• Acceptance above HVN vs rejection = trend confirmation.
• Key institutional price levels for swings.

19. `Pivot Points` · Levels

• Session/day pivot, R1/R2, S1/S2 for intraday.
• 30 / 60 / 240 intraday S/R framework.
• Opening range context for day trades.

20. `Footprint` · Volume

• Intrabar bid/ask aggression and delta-style read.
• Short TF (1/5/15) order-flow confirmation.
• Validate breakout or absorption at level.

21. `News and Fundamentals` · Context

• Earnings, guidance, sector news overlay on chart.
• Catalyst-aware TA when user asks “before/after earnings”.
• Not a substitute for bloomberg-oracle-terminal deep news search.

22. `Volume Weighted Average Price` · Volume

• Intraday fair value anchor on 60 / 240.
• Mean reversion to VWAP in balanced sessions.
• Institutional benchmark for day-trade bias (above/below VWAP).

23. `Break of Structure` · SMC

• BOS / CHoCH for trend continuation or reversal signal.
• Core SMC indicator with Smart-Money-Concept, Multi-TF-BOS-Alignment.
• Structure shift confirmation before entry.

24. `Average True Range` · Volatility

• Stop placement and position sizing by vol.
• Spread vol scoring in multi-symbol benchmark analysis.
• Regime: rising ATR = expansion; falling ATR = compression.

25. `Super Trend` · Trend

• Trailing stop line and trend direction flip.
• Trend-Following and SuperTrend-Ladder-Pullback strategies.
• Simple stay-in-trend vs exit-on-flip rule.

26. `Premium & Discount` · SMC

• Price in upper/lower half of dealing range (premium vs discount).
• SMC bias: long setups in discount, short in premium (with structure).
• Higher-TF filter before taking lower-TF SMC entry.

Starter packs

4.6 Spread ticker format

4.7 search_tickers exchange filter

NASDAQ | NYSE | CRYPTO | NYSE Arca | NYSE American

4.8 Optional fields

5. Common errors and wrong use cases

5.1 Error signals

5.2 Wrong use cases (do not)

5.3 Response handling notes