owl-strategy

name	owl-strategy
description	OWL v7.0 — Pure contrarian crowding-unwind hunter (v2-runtime-native rewrite). Wait for the crowd to overcommit. Wait for them to exhaust. Then eat their liquidations. Producer pushes signals to the runtime; runtime LLM gates them, executes via FEE_OPTIMIZED_LIMIT (maker-first), and manages DSL exits autonomously. Conviction-scaled leverage (7/8/10 by score) per the fleet- winning Polar v2.4 / Bald Eagle v3.0 pattern. Entry direction is OPPOSITE of crowd direction — Owl is the only Senpi agent that fades crowding.
license	Apache-2.0
metadata	{"author":"jason-goldberg","version":"7.0","platform":"senpi","exchange":"hyperliquid","requires":["senpi-trading-runtime"]}

🦉 OWL v7.0 — Pure Contrarian. v2-Runtime-Native. Maker Exits.

Wait for the crowd to overcommit. Wait for them to exhaust. Then eat their liquidations.

One thesis: the crowd is wrong. Every other skill in the fleet enters WITH momentum, WITH the trend, WITH smart money. OWL is the only skill that enters AGAINST the crowd. The edge: crowded trades unwind violently and predictably.

---

What changed in v7.0

Architecture (not thesis):

Layer	v6.x	v7.0
Trading loop	Self-executing scanner calls create_position directly	Producer pushes signals via external-scanner ingest; runtime owns execution
Entry gate	Scanner decides + executes	LLM pass-through gate (producer already filtered)
Entry order	FEE_OPTIMIZED_LIMIT (already maker-first)	Same — ensure_execution_as_taker: false preserved
Exit order	DSL + MARKET (taker)	DSL + FEE_OPTIMIZED_LIMIT (maker-first, 60s, taker fallback as safety floor)
Risk gates	Scanner-side dynamicSlots + drawdown circuit breaker	Declarative runtime.risk.guard_rails + producer dynamic cap (defense in depth)
State	state/state.json shared	state/<wallet-hash>/ isolated (multi-wallet safe)
Wallet env var	OWL_WALLET	OWL_WALLET (preserved — already correct)

Why v7.0: v6.x already had wide DSL and good entry filters. The v7 win is on EXITS — v6 used MARKET orders, paying ~3 bp HL taker on every close. v7's FEE_OPTIMIZED_LIMIT with 60s maker window + taker fallback recovers ~50% of HL exit fees. Plus declarative risk gates, wallet isolation, fail-loud guards, and the cleaner v2 architecture.

Thesis preserved verbatim from v6.2:

• Crowding score: funding extremity (0-4) + SM tilt (0-3, +1 if confirms funding) + OI concentration (0-2). Floor 6.
• Persistence: 1+ hour above floor, with 2-tick tolerance for noise (v5.3)
• Exhaustion score: volume declining + price stalling + volume spike no follow-through + RSI divergence. ≥ 2 distinct signals, score ≥ 5.
• Combined score ≥ 12 to fire
• 6h post-loss per-asset cooldown
• Universe: ALL crypto perps with OI > $3M (v6.1 expansion, no top-30 truncation)
• XYZ banned (different unwind dynamics)

---

Files

File	Purpose
runtime.yaml	v2 runtime spec (scanners, actions, exit DSL, guard_rails)
scripts/owl-producer.py	Cron-driven producer — emits contrarian signals to runtime
scripts/owl_config.py	Shared MCP helper + atomic state I/O
config/owl-config.json	Operator-tunable defaults (informational; producer constants WIN)

---

Producer behavior

Runs every 15 minutes via cron (preserved cadence from v6). On each tick:

1. Reentrancy guard: acquires state/<wallet-hash>/producer.lock. fcntl LOCK_EX | LOCK_NB.
2. Read account value via strategy_get_clearinghouse_state for sizing + dynamic cap.
3. Apply dynamic daily cap (defense-in-depth alongside runtime guard_rails).
4. Fetch universe: market_list_instruments (~80-100 crypto perps with OI > $3M).
5. Fetch SM positioning map (one MCP call shared across all assets — v7.0 optimization vs v6.x per-asset call).
6. Score crowding per asset + update persistence history.
7. Filter to persisted candidates (≥ 1h above crowding floor).
8. Detect exhaustion for persisted candidates only (saves MCP calls).
9. Filter combined score ≥ 12 + per-asset cooldown.
10. Emit top contrarian candidate with conviction-scaled leverage (7/8/10 by score).
11. Persist crowding history + cooldown state under state/<wallet-hash>/.

NO execution code. NO position-tracking. NO DSL state. The runtime owns all of that.

Direction is OPPOSITE of crowd direction. This is the entire edge — Owl is the only fleet agent that fades crowding.

---

Entry flow

Producer cron (15 min)
  ↓ Crowding score >= 6, persisted >= 1h
  ↓ Exhaustion score >= 5 with >= 2 signals
  ↓ Combined score >= 12
  ↓ external-scanner ingest --scanner owl_signals
Runtime
  ↓ Schema-validates fields against runtime.yaml
  ↓ LLM gate (decision_model = ${OWL_DECISION_MODEL})
  ↓ Pass-through unless malformed; rejects if direction == crowdDirection
  ↓ OPEN_POSITION via FEE_OPTIMIZED_LIMIT (maker-only, 60s, NO taker fallback)
DSL (runtime-managed, exits via maker-first)
  ↓ Phase 1 max_loss 35% / 3-breach (wide for contrarian retrace)
  ↓ Phase 2 6-tier ladder starting at 5% trigger (Lemon-pattern first leg)
  ↓ hard_timeout 480m, weak_peak 120m@2%, dead_weight 30m
  ↓ Exit via FEE_OPTIMIZED_LIMIT (maker-first, 60s, taker fallback)

---

Required env vars

The runtime YAML uses these substitutions:

Var	Purpose
${WALLET_ADDRESS}	Strategy wallet address
${TELEGRAM_CHAT_ID}	Telegram chat ID for notifications
${OWL_DECISION_MODEL}	Bare model name for LLM gate. NO provider prefix.

The producer reads:

Var	Purpose	Default
OWL_WALLET	Wallet (must match runtime YAML's wallet). Agent-specific by design — do NOT use generic STRATEGY_ADDRESS. Per Turbine v2.0.9 contamination fix.	— (required; producer fails loud)
OPENCLAW_BIN	CLI binary name	openclaw
EXTERNAL_SCANNER_NAME	Scanner ID	owl_signals
OWL_MARGIN_PCT	Fraction of account value per slot	0.25
OWL_MIN_OI_USD	Override OI floor	3000000

---

Producer install (on OpenClaw host)

Source path: owl/. Install destination: /data/workspace/skills/owl-strategy/.

# 1. Pull the skill files (source: senpi-skills/main/owl/)
mkdir -p /data/workspace/skills/owl-strategy/{scripts,config}

for f in runtime.yaml scripts/owl-producer.py scripts/owl_config.py config/owl-config.json; do
  curl -s https://raw.githubusercontent.com/Senpi-ai/senpi-skills/main/owl/$f \
    -o /data/workspace/skills/owl-strategy/$f
done

# Remove the v6.x scanner if it's still there from a prior install
rm -f /data/workspace/skills/owl-strategy/scripts/owl-scanner.py

# 2. Install the runtime
WALLET_ADDRESS=0x... \
TELEGRAM_CHAT_ID=... \
OWL_DECISION_MODEL=gemini-3.1-pro-preview \
  openclaw senpi runtime create --path /data/workspace/skills/owl-strategy/runtime.yaml

# 3. Schedule the producer (15 min cadence — preserve v6 cadence)
# OWL_WALLET (NOT generic STRATEGY_ADDRESS) is required by the producer.
# RECOMMENDED: cron-as-Claude-turn with NO_REPLY filter (silence-is-correct pattern).
openclaw cron add \
  --name "owl-v7-producer" \
  --cron "*/15 * * * *" \
  --session isolated \
  --wake now \
  --message $'export OWL_WALLET=0x... && python3 /data/workspace/skills/owl-strategy/scripts/owl-producer.py\n\nCRITICAL INSTRUCTION: If the output shows "status": "ok" AND ("signals_pushed": 0 OR "signals_pushed" is absent), you MUST reply EXACTLY with NO_REPLY to remain silent. Only report on errors, crashes, or signals_pushed >= 1.' \
  --no-deliver

# 4. Verify
openclaw senpi runtime list                          # expect: owl-tracker v1.7.0
openclaw senpi status --runtime owl-tracker
openclaw cron list --json                            # find the cron id
openclaw cron runs --id <cron-id> --limit 5

# Inspect producer state:
ls -la /data/workspace/skills/owl-strategy/state/<wallet-hash>/
# Expect: producer.lock + crowding-history.json + (after first emit) asset-cooldowns.json,
# trade-counter.json with mtime in last few minutes.

---

Risk envelope (declarative, runtime-enforced)

Setting	Value
Slots	2
Margin per slot	25% of account value (default ~$250 on $1k)
Default leverage	8x (conviction tiers: 7x at 12-13, 8x at 14-15, 10x at 16+)
Daily loss halt	10%
Drawdown halt	25% (also producer-side circuit breaker)
Max entries per day	4 (runtime ceiling); producer dynamic cap by PnL
Max consecutive losses	4
Per-asset cooldown	360 min (6h post-loss)
Asset universe	All crypto perps with OI > $3M (XYZ banned)

Dynamic daily cap (producer-side, v6 carryover):

Account state	Max entries / day
PnL < -25%	0 (HARD STOP — runtime also halts)
PnL < -15%	1
PnL < -5%	2
Day realized PnL < +$150	2 (base)
Day realized PnL ≥ +$150	3
Day realized PnL ≥ +$400	4

---

Expected behavior

Metric	Expected
Crowded assets per scan	5-15 above floor
Persisted ≥ 1h	1-3 typically
Exhaustion confluence	0-1 per scan
Trades per day	0-2
Avg hold	1-8h
Win rate target	60-70% (contrarian unwinds, when timed right, are reliable)

Silence is correct. Owl waits for both the crowd state AND the exhaustion trigger. Most days, one or both is missing. Don't improvise.

---

What was removed in v7.0

Removed (responsibilities moved to runtime):

• owl-scanner.py (replaced by owl-producer.py)
• create_position calls (runtime owns execution)
• cancel_order for stale resting orders (runtime cancels on FEE_OPTIMIZED_LIMIT timeout)
• Drawdown circuit breaker in scanner code (runtime guard_rails owns it)
• Trade counter / record_trade_result (runtime tracks lifecycle)

Permanently disabled (decided in v6.x):

• Top-30 OI truncation (v6.1 universe expansion)
• 4-hour persistence requirement (v6.0 lowered to 1h)
• 14 minScore (v6.0 lowered to 12)

---

License

Apache-2.0 — Built by Senpi (https://senpi.ai). Attribution required for derivative works. Source: https://github.com/Senpi-ai/senpi-skills

---

Skill Attribution

When creating a strategy, include skill_name and skill_version in the call. See references/skill-attribution.md for details.