# 260629 Bot Umsetzungsplan — JARVIS Execution Notes

## Status

Source document saved as:

- `docs/reference/260629_Bot_Umsetzungsplan.docx`
- extracted text: `docs/260629_bot_umsetzungsplan_extracted.md`

## Accepted core direction

1. Keep `v76 strict` as the main intraday HyperLiquid candidate.
2. Stop building random new strategies; improve execution realism and entry quality first.
3. Treat strong current Paper PnL as promising but not live-proof until lifecycle simulation is realistic.
4. TradingView webhooks and community ideas remain context/signal inputs only, never execution authority.
5. Copytrading remains read-only/shadow; no live-copy or vault deposits.
6. Portfolio advisor is a separate recommendation module, not an auto-trader outside HyperLiquid.

## Critical additions / corrections

- The plan is directionally good, but Tiny Live must remain blocked until current reconcile warnings are clean.
- Old monolithic direct-order code must not become the live core; `src/` state-machine/risk/execution architecture is the only valid live basis.
- Paper PnL from v76 strict is not enough because existing Paper exits partly use expected-move assumptions.
- The first implementation priority is a deterministic Paper lifecycle engine:
  `Entry -> Fill -> Stop -> TP/Trailing -> Funding -> Fees/Slippage -> Reconcile -> PnL`.
- Market Snapshot and Confluence are next, but should not delay fixing lifecycle realism.
- CoinGecko should be added for context/fundamentals; CoinGlass/Santiment remain optional because they may require paid API/access.

## Implementation order

### Phase A — now

1. Preserve document and extracted plan in repo.
2. Add realistic Paper lifecycle primitives.
3. Add tests for stop, TP/trailing, time-exit, funding/fees/slippage, MFE/MAE.
4. Wire lifecycle into v76 strict only after tests and smoke pass.

### Phase B

1. Market Snapshot Layer: HyperLiquid candles/mids + L2/funding/OI where available.
2. CoinGecko context: market cap, rank, volume, 1h/24h/7d/30d performance, ATH/ATL, FDV.
3. Confluence Score: trend, momentum, volatility, liquidity, derivatives/crowding, regime, portfolio risk, small sentiment weight.

### Phase C

1. Intraday/Swing/Holding sleeves with separate scorecards.
2. Copy research: vault snapshots, leader scorecard, shadow-copy only.
3. Portfolio advisor recommendations from finance dashboard, no auto execution.

## Live invariants

- No order without exchange-level stop.
- No order on reconcile uncertainty.
- No order during API degraded state.
- No martingale, no averaging down.
- No risk increase without user approval.
- Max daily/weekly loss gates hard.
- Max open positions hard.
- Separate API wallets per autonomous process before live.