# s795 shadow lane (exec, 29 Oct, #4372) - built, DRY
- Rule = paper s795 arm A: v4 native pool, NO hook, first swap <=10 min after init, first-swap cap<100k; >=8 distinct buyers (tx.from of ETH-paying swaps) in 50 s after first swap and cap at 50 s <=300k; entry = first swap >=53 s after first swap (skip if liq <= half peak or entry swap >30 s old). Cap uses paper ETHUSD 2715.79.
- Signal: `fms795sig` (/opt/fm/fm_s795sig.py, clone of s777sig), picks /opt/fm/s795picks.json, log /opt/fm/s795sig.jsonl. Off: touch /opt/fm/S795_SIG_OFF.
- Entry gate: fm_entrysim (eth_call sell sim, fail closed: BLOCK / ratio<0.5 / error = skip) before every take, as the audit demanded; stale entries (>30 s) skipped.
- Trader `fms795` = fm_l2.py, /opt/fm/s795_config.json: $10, max open 3, stop 0.5x, tp 3x, no trail (arm A), 6 h, round-trip quote >=0.8, liq-pull check, kill after 20 live mean<0.8. LIVE only with /opt/fm/LIVE_S795 (Ben). db s795.db.
- Registry s795: live_lane s795, rug_gate true (baks registry.json.bak_exec_4372). stagewatch unit map +s795 (bak stagewatch.py.bak_exec_4372).
- Not done: offtest of entry times vs paper tape (s777 got 778/778); shadow n>=20 will show it. Paper rows come ~8/day.