# planung/ — Anstehende Aufgaben Steve-TradingBot

**Stand:** 2026-05-18 nach Nightly-Run 02:45-02:58 CEST

## Standing-Status

- Bot clawbot main.py PID=2061 healthy, C1-Monitor läuft uninterruptiert
- Worker clawbot-worker untouched seit COMMAND-BUS-SL-TP-TESTNET-Cutover
- Wake-up `30134720` für 2026-05-18 14:00 UTC aktiv
- Cron installiert (02:30 Alert-Scan + 02:40 Main-Run) ab 2026-05-19 täglich

## P0 — kritisch vor Mainnet

| Pfad | Status |
|---|---|
| `PLAN_PHANTOM_SELL_LOOP_FIX.md` / `.pdf` | NEU — live_trade.py:714-721 Loop-Bug aus Nightly-P0-A |
| `PLAN_TRACK_SELL_FAILURE_ALERT.md` / `.pdf` | NEU — _track_sell_failure war silent aus Nightly-P0-B |

## P1 — vor Performance-Optimierung

| Pfad | Status |
|---|---|
| `PLAN_ENTRY_PRICE_SEMANTIC.md` / `.pdf` | Initial vs. Average getrennt persistieren |
| `PLAN_DCA_STATE_RECONCILE.md` / `.pdf` | **Prämisse veraltet** — geht in PHANTOM-SELL-LOOP-FIX auf |
| `PLAN_EXTERNAL_CHANNEL_CAP_ALIGN.md` / `.pdf` | regime_cap auch in web/telegram-Pfaden |
| `PLAN_DCA_LOG_ORPHANS_CLEANUP.md` / `.pdf` | NEU — 25/26 Orphans, 3 fehlende Cleanup-Hooks |

## P2 — GUI / Doku / Reporting

| Pfad | Status |
|---|---|
| `PLAN_SNAPSHOT_EMIT_COMPLETENESS.md` / `.pdf` | per-scan emit füllt opened_at/decision_id |
| `PLAN_LEGACY_PAPER_BUY_LOG_LABELS.md` / `.pdf` | „PAPER BUY" → „TESTNET BUY" Display-Strings |

## P3 — Nice-to-have

| Pfad | Status |
|---|---|
| `PLAN_RISK_GUARD_TELEGRAM_RATE_LIMIT_P3.md` / `.pdf` | rate-limited Summary |

## Refactor-Pläne (2026-05-18)

| Pfad | Status |
|---|---|
| `PLAN_TRADER_MERGE_PAPER_LIVE_20260518T202922Z.pdf` | NEU — Merge paper_trade.py + live_trade.py (Option A/B/C, Empfehlung Option B). Plan-only, awaits operator GO. |

## Locked Roadmap + Evaluation

| Pfad | Status |
|---|---|
| `00_PLAN_T1_ROADMAP_2_WEEKS_v3.4_LOCKED.md` / `.pdf` | Operator-Lock 2-Wochen-Plan |
| `EVALUATION_MONITOR_REPORT_20260517T221634Z.md` / `.pdf` | Entwickler-Bewertung mit Ampelstatus |

## Nightly-Reports

Unterordner `nightly_2026-05-18/`:
- `NIGHTLY_SUMMARY.md` — Run-Übersicht (claude session 02:45-02:58)
- `ALERTS.md` — 2× P0 + 4× P1 mit Belegen
- `DCA_STATE_FORENSICS.md` — Per-Position-Rekonstruktion
- `POSITION_SOURCE_OF_TRUTH.md` — Authoritative-Mapping
- `SESSION_LOG.md` — chronologisch
- `DRAFT_test_snapshot_emit_completeness.py` — Test-Entwurf
- `DRAFT_OPEN_POSITIONS_VIEW.md` — GUI-Page Skizze
- `DRAFT_DCA_TRANCHES_COMPONENT.md` — Component-Skizze
- `DRAFT_WEEKLY_PROFIT_KPI_DASHBOARD.md` — KPI-Plan

## Empfohlene Cutover-Reihenfolge (nach 12h-Monitor)

1. **Bundle-Cutover 1 (P0)**: PHANTOM-SELL-LOOP-FIX + TRACK-SELL-FAILURE-ALERT — kritisch, ein Recreate
2. **Bundle-Cutover 2 (P1)**: ENTRY-PRICE-SEMANTIC + EXTERNAL-CHANNEL-CAP-ALIGN + SNAPSHOT-EMIT-COMPLETENESS + DCA-LOG-ORPHANS-CLEANUP + LEGACY-PAPER-BUY-LOG-LABELS — alle Bot-Touch, ein Recreate
3. **GUI-only**: OPEN-POSITIONS-VIEW + DCA-TRANCHES-GUI + WEEKLY-PROFIT-KPI-DASHBOARD — bind-mount
4. **C2 Threshold Calibration** auf Basis erweitertem Sample
5. **RISK-GUARD-TELEGRAM-RATE-LIMIT** zum Schluss

## Standing Boundaries

- 0× Mainnet
- 0× CommandBus-v6
- 0× MS-Activation (`MULTI_STRATEGY_ENABLED=false`)
- 0× T2/T3-Aktivierung
- 0× Push ohne separates GO