JobnTop

Get Status

Collector running?, row counts per table, disk usage, and data coverage window (earliest..latest ts per indicator). Reads status.json (no DB lock) + a quick DB scan.

Get Obd

Per-band OBD ratios (perp & spot) at a timeframe (5m/15m/1h/4h/1d) + the latest top levels + perp +/-30 / spot +/-60 thresholds and whether each band crosses them.

Get Indicator Snapshot

Latest values across all seven indicators + Screen-1 bias + decide() output.

Read Market

Course-style multi-timeframe read for mode in {scalp, day, swing}: bias (higher TF) -> confirmation (mid TF) -> execution (lower TF), each with data + the course rule, then the decide() LONG/SHORT/WAIT + confidence.

Run Backtest

Partial backtest over accumulated data (klines+VBS+funding). Flags that OBD/liq history only exists from collection start (D-009). Read-only; results not persisted.

Get Journal

The T-027 performance Journal: recent recorded signals + per-level/per-confidence analytics (win_rate, expectancy = win_rate*avg_win - loss_rate*avg_loss, max_drawdown, worst_losing_streak, sample_size, collection_window, and T-054 raw_signal_count vs unique_setup_count -- distinct setup_cluster_id, so one market event's 3 level lenses count as ONE setup). Pass level scalp/intraday/swing or omit for all. since: 0 (DEFAULT) = full history; -1 = since the 2026-06-04 deploy (D-018, the clean post-c

Get Demo Portfolio

Demo paper-portfolio (T-034) over the recorded Journal: compounds risk ~2% per trade (E9, R/R>=1:2) over CLOSED signals -> equity curve, final balance, win rate, expectancy (R), max drawdown, worst losing streak + per-level summary + a forecast (NOT a promise). `since` default (-1) = headline run since the 2026-06-04 deploy (D-018); since=0 = full history; since=-2 = the T-049 calibration-era clock (D-022; None until set -> full history); since=-3 = the T-052 measurement-integrity era clock (D-0

Get Graduation Readiness

READ-ONLY graduation-readiness surface (T-116): per level (scalp/intraday/swing) x per engine (TRDR primary / Models) a deterministic verdict in {NOT_READY, READY_PENDING_REVIEW, INSUFFICIENT_DATA} over the EXISTING accounting + journal fields (sample vs the 50/25/8 minimums, expectancy_net_r after fees+slippage+funding > a provisional 0.0 bar, the v5 clock_started flag), with the WHY-no-PASS degraded reason, an A/B TRDR-vs-Models side-by-side (NO winner), and the five PENDING reviewer threshold

Study Thresholds

READ-ONLY 15m per-level THRESHOLD-DISCOVERY (T-138): per (indicator, level) the threshold-vs-edge CURVE on the 15m lens for each TRDR indicator + the candidate_repeating gate (n>=8 AND events span >1 UTC day AND the win-rate direction holds pre/post T-121), plus the minute_cov + DEGRADED flag and the pre/post-T-121 split (never a bare win-rate). indicator = obd|vbs|oi|liq|all (default all; 'all' = EXACTLY the frozen four) OR a 039 NON-VOTING descriptive LENS: funding_regime|ls_regime|tl_proximit

Read Window

READ-ONLY window market-autopsy (T-145; D-042 rev 3): six INDEPENDENT provenance-tagged layers over a window -- header + minute_cov + DEGRADED + T-121 era split; 15m swings (k=2) + 1h/4h/1d levels + Models CISD; the 7 TRDR indicators + Price RSI each a SEPARATE piece (value + threshold + crossed? + direction, never merged); deep walls; per-swing continuation at 60/240/1440 MINUTES (right-censored never a loss); repetition over FIXED keys (no mining). Window = from_iso+to_iso OR from_price+to_pri

Count Repetition

READ-ONLY whole-store repetition counter (spec 052; PROGRAM-050 U3). How often has a FIXED pre-declared key condition occurred across the WHOLE available store (NOT a window -- no 35-day clamp)? Keys: TRDR (engine, piece_id, threshold_id, direction, level, era) over the signals journal; Models (formation_type, day_type, level, era) over signals_models. engine = trdr|models (omit for both). key_filter = optional exact-match dict over the DECLARED keys ONLY (e.g. {"level": "scalp", "era": "pre_t12

Data Control

READ-ONLY data control plane (spec 053; PROGRAM-050 U4) -- the backend surface for the estate's data tab. op (closed enum): plan | coverage | status | quarantine_inventory. plan = decompose an inclusive ISO-date range (range_from/range_to) into ISO-week units routed per family (chd_liq -> onbox_fetch; l2_depth -> offbox_stage, the committed zero-network staged work order -- heavy reconstruction NEVER routes on-box; requested_route="onbox" for l2_depth refuses by name citing the 2026-07-29 satura

Strategy Workbench

READ-ONLY strategy workbench (spec 055; PROGRAM-050 U6) -- the registered- candidate registry over the committed claim ledger. op (closed read enum): board | candidate | health. board = all candidates in per-level columns (scalp/intraday/swing) with DERIVED lifecycle states (labels over recorded history, never measurements), per-level concurrency counts vs the 2-3 bound, and NF-001..006 verbatim. candidate = one key ("<level>/<name>") -> the full row history verbatim + the derived state + the un

Lens Value

LAB-7 per-lens incremental value (spec 057) -- READ-ONLY, DESCRIPTIVE, labels-not-signals. Measures the canonical-VALUES baseline trigger family (+/-30 / 1-2.5%, both directions, served SPOT surface, scalp geometry) against baseline+lens arms (vbs, liquidations, oi, funding, obd_context, top_levels -- the charter set; long_short + price_rsi NAMED not-armed) over 048-class chronological splits (the 0.2 in-window tail is NEVER read; the holdout stays LOCKED). Per-event NET = m1-per-event.v1 with s