applies to:
kill:
No findings match the current filter.
Ranked tradeable edges across all findings. Shipped + testing at top, then high-confidence pending.
No edges yet. Findings need confidence ≥ medium to surface here.
UTC hour schedule — click cells to toggle tracks
Current active config (merged with shipped gate/schedule findings):
Latest autonomy report
Recent LLM-initiated changes
No LLM changes logged yet.
Create operator finding
🎼 Conductor (live)
regime: ·
conf: ·
at:
🟡 PROBATION ACTIVE — all mults forced to 1.0 until
firing_multipliers ():
self-review:
risks:
calibration hit rate:
Waiting for first conductor cycle…
💎 Meme Intel (top 5)
No ranking yet.
LLM cost today — breakdown
Guard kill-switch events (last 5)
None fired. Good.
🌑 Shadow Mode — trades that would-have-fired but dice rolled shadow ($1 paper)
Recorded when conductor assigns mult < 0.5 to a rule AND random roll exceeds mult. Preserves learning data from deprioritized rules without risking capital.
Shadow log (wildcard, last 20)
No shadow entries yet — conductor hasn't deprioritized any actively-firing rule yet, or no rule has fired under a deprioritized mult.
Calibration (is conductor right?)
Hit rate:
⚠️ Conductor hit rate below 60% — operator should review decisions
Per-rule:
Loading…
🔬 Methodology Review (twice-daily)
Last run:
auto-applied: ·
queued: ·
vetoed:
✓ auto-applied (SMALL):
:
→
⏳ queued (SIGNIFICANT — operator approval):
:
No methodology run yet.
📡 Data Source Health
Dead sources (documented, not tracked):
👛 Smart-Money Wallet Watchlist
Helius-polled every 10 min. Cluster signals flow into conductor + meme-intel. Auto-discovery available after 10+ winning meme trades via POST /api/discover-from-wins.
Watchlist ():
Watchlist empty — operator must seed via PUT /api/wallet-list or wait for auto-discovery.
Recent Cluster Signals (last 30 min)
No cluster signals yet. Need 2+ watched wallets buying same token within 15 min.
🧪 Wildcard Experimental Rules
Analyst-generated hypotheses shipped as test rules (8% size, short horizon, auto-lifecycle: retire if WR<35% at n=20; scale up at WR>55%).
n= / target
· wins=
· WR=
· pnl=$
No experimental rules yet. Analyst deep-cycle will propose; operator approves via Findings tab → ship-as-wildcard.
Natural-language command
Examples: “approve F-012”, “reject the WC-W1 LSR thing”, “add veto on leverage changes”.
Veto list
Change types autonomy may never auto-ship
Operator notes
Last 72h — everything that happened
Merges audit log, LLM changes, autonomy cycles, guard events.
No recent events.