Add first-hour QTA2 research and persistent position ledger

.claude-plugin/marketplace.json

@@ -29,7 +29,7 @@
       "displayName": "Quant Stock Technical",
       "source": "./plugins/quant-stock-technical",
       "description": "Acquire audited KIS Korean and U.S. EOD data and screen a deterministic KOSPI, KOSDAQ, NYSE, and NASDAQ technical universe.",
-      "version": "0.4.4",
+      "version": "0.4.5",
       "author": {
         "name": "mrcha033"
       },
@@ -137,7 +137,7 @@
       "displayName": "Quant Stock Polling Trader",
       "source": "./plugins/quant-stock-polling-trader",
       "description": "Build and supervise deterministic daily KIS shadow sessions from official calendars through reconciliation.",
-      "version": "0.5.3",
+      "version": "0.5.4",
       "author": {
         "name": "mrcha033"
       },

.github/workflows/validate.yml

@@ -28,6 +28,12 @@ jobs:
         run: python -B scripts/sync_distribution.py --check
       - name: Run deterministic calculator self-test
         run: python -B skills/quant-stock-technical/scripts/analyze_stock.py --self-test
+      - name: Run first-hour research calculator self-test
+        run: python -B skills/quant-stock-technical/scripts/analyze_stock_v2.py --self-test
+      - name: Run read-only opening-hour collector self-test
+        run: python -B skills/quant-stock-technical/scripts/collect_open1h_research.py --self-test
+      - name: Run opening-hour evaluator self-test
+        run: python -B skills/quant-stock-technical/scripts/evaluate_open1h_research.py --self-test
       - name: Run four-exchange universe builder self-test
         run: python -B skills/quant-stock-technical/scripts/build_universe_manifest.py --self-test
       - name: Run deterministic universe screen self-test
@@ -42,6 +48,8 @@ jobs:
         run: python -B skills/quant-stock-polling-trader/scripts/market_calendar.py self-test
       - name: Run recurring daily pipeline self-test
         run: python -B skills/quant-stock-polling-trader/scripts/daily_pipeline.py self-test
+      - name: Run strategy-position lifecycle self-test
+        run: python -B skills/quant-stock-polling-trader/scripts/position_lifecycle.py --self-test
       - name: Run systemd generator self-test
         run: python -B skills/quant-stock-polling-trader/scripts/systemd_units.py self-test
       - name: Run mandatory handoff-gate self-test
@@ -76,6 +84,12 @@ jobs:
         run: python -B tests/test_screen_universe.py
       - name: Run screening and execution regression
         run: python -B tests/test_quant_execution.py
+      - name: Run QTA2 execution and market-regime regression
+        run: |
+          python -B tests/test_qta2_execution.py
+          python -B tests/test_qta2_market_regime.py
+      - name: Run strategy-position ledger regression
+        run: python -B tests/test_position_lifecycle.py
       - name: Run systemd unit generator regression
         run: python -B tests/test_systemd_units.py
       - name: Run KIS U.S. EOD regression

plugins/quant-stock-polling-trader/.claude-plugin/plugin.json

@@ -2,7 +2,7 @@
   "$schema": "https://json.schemastore.org/claude-code-plugin-manifest.json",
   "name": "quant-stock-polling-trader",
   "displayName": "Quant Stock Polling Trader",
-  "version": "0.5.3",
+  "version": "0.5.4",
   "description": "Run a direct deterministic daily KIS shadow workflow from official calendars through reconciliation.",
   "author": {
     "name": "mrcha033",

plugins/quant-stock-polling-trader/.codex-plugin/plugin.json

@@ -1,6 +1,6 @@
 {
   "name": "quant-stock-polling-trader",
-  "version": "0.5.3",
+  "version": "0.5.4",
   "description": "Run a direct deterministic daily KIS shadow workflow.",
   "author": {
     "name": "mrcha033",

plugins/quant-stock-polling-trader/skills/quant-stock-polling-trader/SKILL.md

@@ -88,6 +88,12 @@ without making a Korean session wait for U.S. EOD or vice versa.
 Finish an initial multi-year bootstrap before relying on a daily timer; the
 timer is intended to resume and increment an existing cache.
 
+Set optional `qta_method_version` to `qta-2.0.0` for the research-only
+first-hour model. The runner then performs one read-only benchmark-regime call
+per active exchange and cycle before stock quotes. A negative, stale,
+wrong-session, or unavailable regime observation blocks new entries for that
+cycle. QTA2 cannot run in paper or live mode.
+
 To receive optional ntfy completion/block notifications, place these values in
 the same mode-`0600` environment file used by the generated services:
 
@@ -111,6 +117,15 @@ The account alias is nonsecret and configured by the user. A current session
 is never entered twice. An unresolved broker mutation from a paper-mode run
 still requires reconciliation before another entry.
 
+Filled strategy quantities are stored in
+`RUNTIME_ROOT/state/<account-alias>/<market>/positions.sqlite3`, outside the
+date directory. Before a later session reads quotes, the runner reconciles
+that ledger against the frozen KIS account snapshot. Matching positions roll
+forward as `CARRY_EXIT_ONLY`; quantity differences, working orders, duplicate
+ownership, or a prior `MANUAL_BLOCK` stop new entries. Holdings not created by
+this strategy remain visible as unmanaged broker positions and are never
+silently adopted or sold.
+
 ## Direct planning and polling
 
 For a non-recurring session, read `references/execution-contract.md`,
@@ -170,6 +185,7 @@ python3 scripts/account_snapshot.py self-test
 python3 scripts/market_calendar.py self-test
 python3 scripts/daily_pipeline.py self-test
 python3 scripts/plan_orders.py --self-test
+python3 scripts/position_lifecycle.py --self-test
 python3 scripts/run_session.py self-test
 python3 scripts/reconcile.py --self-test
 python3 scripts/systemd_units.py self-test

plugins/quant-stock-polling-trader/skills/quant-stock-polling-trader/references/daily-pipeline-contract.md

@@ -38,6 +38,7 @@ Use `qta-daily-shadow-config/v2`:
   "mode": "shadow",
   "history_start_date": "2021-01-01",
   "minimum_sessions": 756,
+  "qta_method_version": "qta-2.0.0",
   "request_interval_ms": 120,
   "catalog_coverage_contract": {
     "schema": "qta-catalog-coverage-contract/v1",
@@ -94,7 +95,11 @@ Use `qta-daily-shadow-config/v2`:
 }
 ```
 
-Only the fields shown above are required. Two optional fields may be added:
+`qta_method_version` is optional and defaults to `qta-1.0.0`. Select
+`qta-2.0.0` only for shadow research; the screen and runner then enforce its
+liquidity and same-session market-regime contracts.
+
+Three other optional fields may be added:
 
 ```json
 {
@@ -118,6 +123,12 @@ The intent ledger is:
 RUNTIME_ROOT/state/<account-alias>/<market>/<session-date>
 ```
 
+The strategy-position ledger persists across dates at:
+
+```text
+RUNTIME_ROOT/state/<account-alias>/<market>/positions.sqlite3
+```
+
 The workflow descriptor is
 `RUNTIME_ROOT/current/<market>.json`, and date-specific artifacts are under
 `RUNTIME_ROOT/workflows/<market>/<session-date>`.

plugins/quant-stock-polling-trader/skills/quant-stock-polling-trader/references/execution-contract.md

@@ -5,7 +5,7 @@
 Keep three independent contracts:
 
 ```text
-qta-1.0.0          completed T-1 adjusted daily data -> technical payload
+qta-1.0.0/2.0.0    completed T-1 adjusted daily data -> technical payload
 qta-screen-1.0.0   legacy KR/US manifest -> deterministic candidates
 qta-screen-1.1.0   selected-market manifest -> exchange-ranked candidates
 open1h-exec-2.0.0  broker/account snapshots -> intents and order lifecycle
@@ -62,6 +62,15 @@ resolved-tick alignment, and the emitted 2R take-profit geometry. A smaller,
 reranked, or semantically forged object cannot make itself acceptable merely
 by hashing its own contents.
 
+For `qta-2.0.0`, the validator instead recomputes its first-hour total,
+verifies the market-specific turnover floor, and requires the exact
+research-only market-regime policy. QTA2 can run only in shadow mode. Before
+stock quotes in each cycle, the runner reads KOSPI Composite, KOSDAQ
+Composite, Nasdaq Composite, or the documented S&P 500 broad-market proxy for
+NYSE candidates. It verifies the current-session timestamp, a maximum age of
+90 seconds, and return from previous close of at least zero basis points. A
+failed check skips new entries for that exchange and cycle.
+
 Each account position must include `market` and `symbol`, plus `exchange` for a
 v2 screen. Each item in `open_orders` must represent a currently working order
 and include `side, market, symbol`, plus `exchange` for v2. Existing positions
@@ -164,6 +173,9 @@ KIS live US shadow: ceil(intent_count * 1 * 0.12s)
 Toss shadow: ceil(intent_count * 2 * 0.10s)
 ```
 
+QTA2 adds one KIS benchmark request per active exchange to the same cycle
+budget.
+
 Reject a configured interval below that floor. This is only an API-pacing
 feasibility bound, not proof of network latency. Production promotion also
 requires observed shadow percentiles.
@@ -211,6 +223,16 @@ At the window end, freeze new entries, cancel outstanding paper-mode entry
 orders, and reconcile. Continue managing already filled paper positions.
 Automatic market flattening is not implemented.
 
+The account/market strategy-position ledger is stored one directory above the
+date-specific intent ledger. Authoritative cumulative entry fills create or
+update a position with its original stop, target, tick, FX, and plan identity.
+Before the next session, exact broker quantities are reconciled and valid
+residuals roll to `CARRY_EXIT_ONLY`; they are excluded from new additions.
+The bundled lifecycle module deterministically calculates stop, gap-stop,
+target, close-liquidation, partial-fill, and daily-loss decisions, but the
+current polling service does not yet send live exit mutations. Do not describe
+that decision module as active automatic liquidation.
+
 ## State machine
 
 ```text