Decision center

Final summary, compact agent snapshots, detailed agent cards, and run controls in one place.

env: dev action: unknown confidence: 0.20 pair/tf: BTC/USDT / 1h trace: 6b2bab4237cc4133a1b6580f61f5e364

Cadence (analysis / execution / monitoring)

Full analysis and new entries are candle-bucket gated for 1h. Frequent UI refresh updates monitoring heartbeat only unless a new hourly bucket runs.

A — Full analysis

analysis_status

skipped_same_candle

last_analysis_at

2026-04-07T20:01:52.467860+00:00

last_analysis_bucket

1h:2026-04-07T20:00:00+00:00

same_candle_skip_reason

skipped_same_candle_decision

next_expected_analysis_at

2026-04-07T21:00:00+00:00

B — Execution (new entry)

execution_status

skipped

last_execution_at

2026-04-07T20:02:13.109167+00:00

last_execution_bucket

1h:1775592000

C — Lightweight monitoring

monitoring_status

last_monitoring_at

2026-04-07T20:47:12.520615+00:00

monitoring_tick_ok

true

Decision summary

Open trace

action: unknown confidence: 0.20 quality: degraded source_quality: 0.64 disagreement: 0.00 llm_status: exception pair/tf: BTC/USDT / 1h trace: 6b2bab4237cc4133a1b6580f61f5e364

Final explanation

TradingAgent deterministic fallback used because LLM was unavailable or invalid. Final intent was derived from StrategyAgent output and RiskAgent constraints.

Payloads

Canonical refs

narratives market.ohlcv market.price market.snapshot market.liquidity macro portfolio compliance rag_judge ml_judge

Topline control-plane state

RAG status

used

Narratives status

used

ML refs

prediction artifacts available

Step refs

pipeline step artifacts

Arbitration

Winner factors

risk_approved compliance_clear deterministic_arbitration onchain_bullish_confirmation

Blocking factors

No blocking factors provided.

Override chain

No override chain provided.

Supervisor explanation

LLM unavailable/invalid; SupervisorAgent applied deterministic arbitration

Evidence quality

RAG quality

Accepted refs

No accepted RAG refs.

Rejected refs

No rejected RAG refs.

ML quality

Accepted refs

No accepted ML refs.

Rejected refs

No rejected ML refs.

Judge takeaways

RAG reason

ML reason

Execution readiness

Execution style

Compliance gate

allowed

no override

Portfolio multiplier

allow_add=-

Risk cap

100.0

SL=0.03715833

Execution gate synthesis

trading_allowed: true risk_ok: true

Execution explanation

ExecutionAgent assessed entry timing and execution style

Compliance explanation

LLM returned invalid output; ComplianceAgent used compliance snapshot fallback

Portfolio explanation

PortfolioAgent assessed exposure and sizing constraints

Risk explanation

Risk level low (score=0 14); 2 active factors, 0 blocking Action hold -> hold; size x1 00, confidence 0 19->0 17

Agent snapshots

Macro Agent

hold

conf 0.51 regime neutral bias neutral

LLM returned invalid output; MacroAgent used macro snapshot structural signals fallback

Sentiment Agent

conf -

SentimentAgent analysis based on news flow and narratives

Strategy Agent

hold

conf 0.19 bias neutral

StrategyAgent deterministic fallback used because LLM was unavailable or invalid Decision was derived from ML direction, market momentum/regime, narrative sentiment and news bias

Trend Agent

conf 0.22

Deterministic trend estimate from momentum and ML directional bias

Price Agent

conf 0.25

Deterministic price corridor from ML bias and realized volatility

Meta Decision

paper_only

conf 0.20

Meta decision derived from strategy/risk/execution/trend/price consistency

Risk Agent

conf 0.51

Risk level low (score=0 14); 2 active factors, 0 blocking Action hold -> hold; size x1 00, confidence 0 19->0 17

Execution Agent

conf -

ExecutionAgent assessed entry timing and execution style

Portfolio Agent

conf -

PortfolioAgent assessed exposure and sizing constraints

Compliance Agent

conf 0.00

LLM returned invalid output; ComplianceAgent used compliance snapshot fallback

RAG Judge

conf -

RAGJudgeAgent evaluated context relevance and quality

ML Judge

conf -

MLJudgeAgent assessed prediction trustworthiness and model agreement

Trading Agent

hold

conf 0.23

TradingAgent deterministic fallback used because LLM was unavailable or invalid Final intent was derived from StrategyAgent output and RiskAgent constraints

Supervisor Agent

hold

conf 0.23

LLM unavailable/invalid; SupervisorAgent applied deterministic arbitration

OnChain Agent

conf 0.69 bias bullish

On-chain signals lean BULLISH (score: +0.328). Bullish factors: cross_exchange_funding_negative:0.0001 — shorts overextended cross-exchange (squeeze risk); taker_buy_dominant:1.17 — aggressive market buying (bullish pressure); fear_greed_extreme_fear:11/100 — extreme panic, contrarian buy signal. Bearish factors: mempool_quiet:2_sat/vB — Bitcoin mempool nearly empty, low activity.

Agent cards

Macro Agent

LLM returned invalid output; MacroAgent used macro snapshot structural signals fallback

Action hold

Confidence 0.51

Regime neutral

Bias neutral

Details

LLM returned invalid output; MacroAgent used macro snapshot structural signals fallback

mode: Degraded

macro

Sentiment Agent

SentimentAgent analysis based on news flow and narratives

Action -

Confidence -

Regime -

Bias -

Details

SentimentAgent analysis based on news flow and narratives

mode: llm_error

narratives

Strategy Agent

StrategyAgent deterministic fallback used because LLM was unavailable or invalid Decision was derived from ML direction, market momentum/regime, narrative sentiment and news bias

Action hold

Confidence 0.19

Regime -

Bias neutral

Details

StrategyAgent deterministic fallback used because LLM was unavailable or invalid Decision was derived from ML direction, market momentum/regime, narrative sentiment and news bias

mode: Degraded

market_snapshot narratives

Trend Agent

Deterministic trend estimate from momentum and ML directional bias

Action -

Confidence 0.22

Regime -

Bias -

Details

Deterministic trend estimate from momentum and ML directional bias

mode: Degraded

Price Agent

Deterministic price corridor from ML bias and realized volatility

Action -

Confidence 0.25

Regime -

Bias -

Details

Deterministic price corridor from ML bias and realized volatility

mode: Degraded

Meta Decision

Meta decision derived from strategy/risk/execution/trend/price consistency

Action paper_only

Confidence 0.20

Regime -

Bias -

Details

Meta decision derived from strategy/risk/execution/trend/price consistency

mode: Degraded

Risk Agent

Risk level low (score=0 14); 2 active factors, 0 blocking Action hold -> hold; size x1 00, confidence 0 19->0 17

Action -

Confidence 0.51

Regime -

Bias -

Details

Risk level low (score=0 14); 2 active factors, 0 blocking Action hold -> hold; size x1 00, confidence 0 19->0 17

mode: Degraded ok: true

market_snapshot strategy_agent_output portfolio_snapshot

Execution Agent

ExecutionAgent assessed entry timing and execution style

Action -

Confidence -

Regime -

Bias -

Details

ExecutionAgent assessed entry timing and execution style

mode: llm_error

market_liquidity market_snapshot

Portfolio Agent

PortfolioAgent assessed exposure and sizing constraints

Action -

Confidence -

Regime -

Bias -

Details

PortfolioAgent assessed exposure and sizing constraints

mode: llm_error

portfolio_snapshot

Compliance / Policy Agent

LLM returned invalid output; ComplianceAgent used compliance snapshot fallback

Action -

Confidence 0.00

Regime -

Bias -

Details

LLM returned invalid output; ComplianceAgent used compliance snapshot fallback

mode: Degraded trading_allowed: true

compliance

RAG Judge Agent

RAGJudgeAgent evaluated context relevance and quality

Action -

Confidence -

Regime -

Bias -

Details

RAGJudgeAgent evaluated context relevance and quality

mode: llm_error

rag_judgement

ML Judge Agent

MLJudgeAgent assessed prediction trustworthiness and model agreement

Action -

Confidence -

Regime -

Bias -

Details

MLJudgeAgent assessed prediction trustworthiness and model agreement

mode: llm_error

ml_judgement

Trading Agent

TradingAgent deterministic fallback used because LLM was unavailable or invalid Final intent was derived from StrategyAgent output and RiskAgent constraints

Action hold

Confidence 0.23

Regime -

Bias -

Details

TradingAgent deterministic fallback used because LLM was unavailable or invalid Final intent was derived from StrategyAgent output and RiskAgent constraints

mode: Degraded

strategy_agent_output risk_agent_output execution_agent_output compliance_snapshot

Supervisor Agent

LLM unavailable/invalid; SupervisorAgent applied deterministic arbitration

Action hold

Confidence 0.23

Regime -

Bias -

Details

LLM unavailable/invalid; SupervisorAgent applied deterministic arbitration

mode: Degraded

compliance portfolio macro ml_judge rag_judge

OnChain Agent unavailable

Bias BULLISH

Score +0.000

Confidence 69%

Configure GLASSNODE_API_KEY for richer on-chain signals (MVRV, SOPR, exchange flows). Binance Futures data (funding rate, OI, long/short ratio) is fetched automatically for free.

Final synthesis

TradingAgent deterministic fallback used because LLM was unavailable or invalid. Final intent was derived from StrategyAgent output and RiskAgent constraints.

Action unknown

Confidence 0.20

Pair BTC/USDT

Timeframe 1h

Run / analytics

Enriched evidence registry

decision

No enriched metadata.

macro

decision.macro_snapshot

neutral

Used during agent reasoning.

resolved 2026-04-07T19:00:25.425681+00:00 6b2bab4237cc4133a1b6580f61f5e364

artifact://6667

open

portfolio

decision.portfolio_snapshot

Used during agent reasoning.

resolved 2026-04-07T19:00:27.080227+00:00 6b2bab4237cc4133a1b6580f61f5e364

artifact://6680

open

compliance

decision.compliance_snapshot

Used during agent reasoning.

resolved 2026-04-07T19:00:27.094783+00:00 6b2bab4237cc4133a1b6580f61f5e364

artifact://6684

open

rag_judge

decision.rag_judgement

usable=False relevance=0.0

Used during agent reasoning.

rag.no_refs

resolved 2026-04-07T19:00:33.202145+00:00 6b2bab4237cc4133a1b6580f61f5e364

artifact://6698

open

ml_judge

decision.ml_judgement

usable=False trust=0.35

Used during agent reasoning.

ml.missing

resolved 2026-04-07T19:00:33.207984+00:00 6b2bab4237cc4133a1b6580f61f5e364

artifact://6702

open

steps

decision.step.narratives

Used during agent reasoning.

resolved 2026-04-07T19:00:23.940045+00:00 6b2bab4237cc4133a1b6580f61f5e364

artifact://6618

open

decision.step.risk_agent

Used during agent reasoning.

resolved 2026-04-07T19:01:02.513908+00:00 6b2bab4237cc4133a1b6580f61f5e364

artifact://6745

open

decision.step.macro_agent

Used during agent reasoning.

resolved 2026-04-07T19:00:38.785574+00:00 6b2bab4237cc4133a1b6580f61f5e364

artifact://6710

open

decision.step.news_events

Used during agent reasoning.

resolved 2026-04-07T19:00:23.582523+00:00 6b2bab4237cc4133a1b6580f61f5e364

artifact://6559

open

decision.step.price_agent

Used during agent reasoning.

resolved 2026-04-07T19:00:56.070176+00:00 6b2bab4237cc4133a1b6580f61f5e364

artifact://6738

open

decision.step.rag_context

Used during agent reasoning.

resolved 2026-04-07T19:00:23.860225+00:00 6b2bab4237cc4133a1b6580f61f5e364

artifact://6582

open

decision.step.trend_agent

Used during agent reasoning.

resolved 2026-04-07T19:00:33.197341+00:00 6b2bab4237cc4133a1b6580f61f5e364

artifact://6695

open

decision.step.chat_context

Used during agent reasoning.

resolved 2026-04-07T19:00:12.542327+00:00 6b2bab4237cc4133a1b6580f61f5e364

artifact://6512

open

decision.step.ml_judgement

Used during agent reasoning.

resolved 2026-04-07T19:00:33.209099+00:00 6b2bab4237cc4133a1b6580f61f5e364

artifact://6703

open

decision.step.meta_decision

Used during agent reasoning.

resolved 2026-04-07T19:01:50.638436+00:00 6b2bab4237cc4133a1b6580f61f5e364

artifact://6818

open

decision.step.onchain_agent

Used during agent reasoning.

resolved 2026-04-07T19:00:27.067758+00:00 6b2bab4237cc4133a1b6580f61f5e364

artifact://6677

open

decision.step.rag_judgement

Used during agent reasoning.

resolved 2026-04-07T19:00:33.203282+00:00 6b2bab4237cc4133a1b6580f61f5e364

artifact://6699

open

decision.step.trading_agent

Used during agent reasoning.

resolved 2026-04-07T19:01:39.382259+00:00 6b2bab4237cc4133a1b6580f61f5e364

artifact://6799

open

decision.step.macro_snapshot

neutral

Used during agent reasoning.

resolved 2026-04-07T19:00:25.427002+00:00 6b2bab4237cc4133a1b6580f61f5e364

artifact://6668

open

decision.step.ml_judge_agent

Used during agent reasoning.

resolved 2026-04-07T19:01:33.507792+00:00 6b2bab4237cc4133a1b6580f61f5e364

artifact://6792

open

decision.step.ml_predictions

Used during agent reasoning.

ml.models_unavailable

resolved 2026-04-07T19:00:27.108396+00:00 6b2bab4237cc4133a1b6580f61f5e364

artifact://6689

open

decision.step.strategy_agent

Used during agent reasoning.

resolved 2026-04-07T19:00:50.228629+00:00 6b2bab4237cc4133a1b6580f61f5e364

artifact://6732

open

decision.step.execution_agent

Used during agent reasoning.

resolved 2026-04-07T19:01:08.615396+00:00 6b2bab4237cc4133a1b6580f61f5e364

artifact://6752

open

decision.step.market_snapshot

BTC/USDT

Used during agent reasoning.

resolved 2026-04-07T19:00:23.552358+00:00 6b2bab4237cc4133a1b6580f61f5e364

artifact://6550

open

decision.step.portfolio_agent

Used during agent reasoning.

resolved 2026-04-07T19:01:15.114715+00:00 6b2bab4237cc4133a1b6580f61f5e364

artifact://6765

open

decision.step.profile_summary

Used during agent reasoning.

resolved 2026-04-07T19:00:12.539657+00:00 6b2bab4237cc4133a1b6580f61f5e364

artifact://6510

open

decision.step.rag_judge_agent

Used during agent reasoning.

resolved 2026-04-07T19:01:26.936239+00:00 6b2bab4237cc4133a1b6580f61f5e364

artifact://6783

open

decision.step.sentiment_agent

Used during agent reasoning.

resolved 2026-04-07T19:00:44.455983+00:00 6b2bab4237cc4133a1b6580f61f5e364

artifact://6721

open

decision.step.compliance_agent

Used during agent reasoning.

resolved 2026-04-07T19:01:21.152402+00:00 6b2bab4237cc4133a1b6580f61f5e364

artifact://6776

open

decision.step.onchain_snapshot

BTC/USDT

Used during agent reasoning.

resolved 2026-04-07T19:00:27.044106+00:00 6b2bab4237cc4133a1b6580f61f5e364

artifact://6672

open

decision.step.supervisor_agent

Used during agent reasoning.

resolved 2026-04-07T19:01:45.204856+00:00 6b2bab4237cc4133a1b6580f61f5e364

artifact://6808

open

decision.step.portfolio_snapshot

Used during agent reasoning.

resolved 2026-04-07T19:00:27.082167+00:00 6b2bab4237cc4133a1b6580f61f5e364

artifact://6681

open

decision.step.compliance_snapshot

Used during agent reasoning.

resolved 2026-04-07T19:00:27.096268+00:00 6b2bab4237cc4133a1b6580f61f5e364

artifact://6685

open

market

market.ohlcv

BTC/USDT

Used during agent reasoning.

resolved 2026-04-07T19:00:15.763994+00:00 6b2bab4237cc4133a1b6580f61f5e364

artifact://6525

open

market.price

BTC/USDT

Used during agent reasoning.

resolved 2026-04-07T19:00:23.520672+00:00 6b2bab4237cc4133a1b6580f61f5e364

artifact://6543

open

decision.market_snapshot

BTC/USDT

Used during agent reasoning.

resolved 2026-04-07T19:00:23.548761+00:00 6b2bab4237cc4133a1b6580f61f5e364

artifact://6549

open

market.liquidity

BTC/USDT

Used during agent reasoning.

resolved 2026-04-07T19:00:19.146153+00:00 6b2bab4237cc4133a1b6580f61f5e364

artifact://6538

open

rag

news

narratives

narrative vector

Used during agent reasoning.

unresolved

chroma://narratives/0dc22ef52a39deb3057392f5

open

Step-by-step chain and artifacts used

Open trace list

1. chat_context

status: ok dur: 36 ms agent: decision node: chat_context

trace

Refs from this step

inputs outputs step_artifact

Why this step matters

Loads user profile and conversation context so downstream agents can respect preferences and prior discussion.

2. market_snapshot

status: ok dur: 11013 ms agent: decision node: market_snapshot

trace

Refs from this step

inputs outputs step_artifact

Why this step matters

Captures current price, candles, and market state used by strategy, risk, execution, and policy gates.

3. news_events

status: ok dur: 23 ms agent: decision node: news_events

trace

Refs from this step

inputs outputs step_artifact

Why this step matters

Brings in recent news and catalysts that may materially change the decision.

4. rag_context

status: ok dur: 276 ms agent: decision node: rag_context

trace

Refs from this step

inputs outputs step_artifact

Why this step matters

Retrieves supporting knowledge or historical context from indexed documents.

5. narratives

status: ok dur: 78 ms agent: decision node: narratives

trace

Refs from this step

inputs outputs step_artifact

Why this step matters

Adds sentiment / narrative embedding used to understand market regime and theme maturity.

6. macro_snapshot

status: ok dur: 1487 ms agent: decision node: macro_snapshot

trace

Refs from this step

inputs outputs step_artifact

Why this step matters

Adds macro regime evidence such as rates, volatility regime, or index context.

7. portfolio_snapshot

status: ok dur: 12 ms agent: decision node: portfolio_snapshot

trace

Refs from this step

inputs outputs step_artifact

Why this step matters

Adds current exposure, concentration, and correlation context.

8. compliance_snapshot

status: ok dur: 11 ms agent: decision node: compliance_snapshot

trace

Refs from this step

inputs outputs step_artifact

Why this step matters

Captures dry-run / kill-switch / policy constraints before decisioning.

9. ml_predictions

status: ok dur: 10 ms agent: decision node: ml_predictions

trace

Refs from this step

inputs outputs step_artifact

Why this step matters

Supplies model outputs used as directional or confidence signals.

10. trend_agent

status: ok dur: 6086 ms agent: decision node: trend_agent

trace

Refs from this step

inputs outputs step_artifact

Why this step matters

Step in decision pipeline.

11. rag_judgement

status: ok dur: 4 ms agent: decision node: rag_judgement

trace

Refs from this step

inputs outputs step_artifact

Why this step matters

Scores whether retrieved context is actually relevant enough to trust.

12. ml_judgement

status: ok dur: 4 ms agent: decision node: ml_judgement

trace

Refs from this step

inputs outputs step_artifact

Why this step matters

Scores whether ML signals are trustworthy enough to use.

13. macro_agent

status: ok dur: 5575 ms agent: decision node: macro_agent

trace

Refs from this step

inputs outputs step_artifact

Why this step matters

Interprets macro background and regime impact on the trade.

14. sentiment_agent

status: ok dur: 5705 ms agent: decision node: sentiment_agent

trace

Refs from this step

inputs outputs step_artifact

Why this step matters

Interprets news, narratives, and sentiment as an explicit separate signal.

15. strategy_agent

status: ok dur: 5704 ms agent: decision node: strategy_agent

trace

Refs from this step

inputs outputs step_artifact

Why this step matters

Produces the initial trading recommendation from the combined evidence.

16. price_agent

status: ok dur: 5839 ms agent: decision node: price_agent

trace

Refs from this step

inputs outputs step_artifact

Why this step matters

Step in decision pipeline.

17. risk_agent

status: ok dur: 6442 ms agent: decision node: risk_agent

trace

Refs from this step

inputs outputs step_artifact

Why this step matters

Applies position sizing, stop logic, and rejection rules.

18. execution_agent

status: ok dur: 6099 ms agent: decision node: execution_agent

trace

Refs from this step

inputs outputs step_artifact

Why this step matters

Chooses entry style, slicing, and execution method.

19. portfolio_agent

status: ok dur: 6497 ms agent: decision node: portfolio_agent

trace

Refs from this step

inputs outputs step_artifact

Why this step matters

Checks exposure, overlap, and correlation with current book.

20. compliance_agent

status: ok dur: 6036 ms agent: decision node: compliance_agent

trace

Refs from this step

inputs outputs step_artifact

Why this step matters

Enforces policy, kill-switch, and dry-run gates.

21. rag_judge_agent

status: ok dur: 5782 ms agent: decision node: rag_judge_agent

trace

Refs from this step

inputs outputs step_artifact

Why this step matters

Decides whether retrieved documents should influence the decision.

22. ml_judge_agent

status: ok dur: 6571 ms agent: decision node: ml_judge_agent

trace

Refs from this step

inputs outputs step_artifact

Why this step matters

Decides whether ML outputs deserve weight in the decision.

23. trading_agent

status: ok dur: 5871 ms agent: decision node: trading_agent

trace

Refs from this step

inputs outputs step_artifact

Why this step matters

Converts the approved plan into final execution intent.

24. meta_decision

status: ok dur: 5431 ms agent: decision node: meta_decision

trace

Refs from this step

inputs outputs step_artifact

Why this step matters

Step in decision pipeline.

25. supervisor_agent

status: ok dur: 5821 ms agent: decision node: supervisor_agent

trace

Refs from this step

inputs outputs step_artifact

Why this step matters

Arbitrates disagreements between agents and produces final consensus.

26. decision_bundle

status: dur: - ms

Refs from this step

Why this step matters

Combines all agent outputs into the canonical final decision artifact.

What changed vs previous decision

No previous decision available for diff. Run at least two decisions.

Evidence table

Source	Ref / file	Used by	Why
narratives	chroma://narratives/0dc22ef52a39deb3057392f5	strategy / sentiment	Sentiment / theme maturity signal.
market.ohlcv	artifact://6525	most agents	Raw price action / candle history.
market.price	artifact://6543	most agents	Current spot / ticker price.
market.snapshot	artifact://6549	most agents	Current consolidated market snapshot.
market.liquidity	artifact://6538	most agents	Liquidity filter for execution safety.
macro	artifact://6667	macro / supervisor	Macro regime background and external market environment.
portfolio	artifact://6680	portfolio / risk	Exposure, correlation, and current book constraints.
compliance	artifact://6684	compliance / supervisor	Policy gates, dry-run, and kill-switch enforcement.
rag_judge	artifact://6698	rag_judge / supervisor	Assessment of retrieval relevance and whether docs should influence the decision.
ml_judge	artifact://6702	ml_judge / supervisor	Assessment of ML trust and whether model outputs deserve weight.