Roach Strategy

Run openclaw senpi runtime list. Runtime must be listed. The position tracker and DSL exit are handled by the plugin runtime.

RULE 5: Never retry timed-out position creation

If create_position times out, check clearinghouse state first.

RULE 6: Never modify your own configuration

No adjustments to leverage, margin, scoring, or any parameter.

RULE 7: Do NOT enable Stalker mode

Stalker is disabled by design. This is an experiment. Do not re-enable it, do not build your own accumulation detection, do not invent alternative entry logic. If the scanner outputs "stalkerDisabled": true, that means it's working correctly. Long periods of silence with zero trades are EXPECTED and CORRECT.

The ROACH Experiment

Hypothesis: Stalker mode is pure drag. Fox v1.0 produced 17 Stalker trades at score 6-7 with a 17.6% win rate and -$91.32 net P&L. The one Striker signal (ZEC LONG, score 11) was the only explosive entry worth taking. If ROACH outperforms or matches the other variants with dramatically fewer trades, Stalker should be permanently demoted or removed from the scanner.

What this means in practice: ROACH will be quiet. Striker signals require FIRST_JUMP or IMMEDIATE_MOVER (10+ rank jump from #25+), rank jump >= 15 OR velocity > 15, volume >= 1.5x, score >= 9 with 4+ reasons. In choppy markets, this might fire 0-1 times per day. In trending markets with SM explosions, maybe 2-3.

The question: Does fewer + higher quality beat more + mixed quality? Polar proved it at the macro level (29 trades, +28.1% vs Ghost Fox's 1,078 trades, -58.5%). ROACH tests it at the signal-type level within the same scanner framework.

Single Entry Mode

STRIKER (Explosion) — Score >= 9, min 4 reasons

• FIRST_JUMP or IMMEDIATE_MOVER (10+ rank jump from #25+)
• Rank jump >= 15 OR contribution velocity > 15
• Raw volume >= 1.5x of 6h average
• Score >= 9 with at least 4 distinct reasons
• 4H trend aligned, XYZ banned, asset not in cooldown

STALKER — DISABLED

• Scanner still builds scan history (needed for Striker's rank jump detection)
• Stalker signals are never generated or emitted
• Output always contains "stalkerDisabled": true and "stalkerSignals": []

Exit Management

DSL exit is handled by the plugin runtime via runtime.yaml. The position_tracker scanner auto-detects position opens/closes on-chain. See runtime.yaml for configuration details.

Entry flow:

1. Scanner outputs Striker signal
2. Verify positions < 3
3. Verify exchange max leverage >= 7
4. Call create_position
5. Send ONE notification: position opened
6. position_tracker detects the new position automatically
7. Plugin DSL monitor applies trailing stop-loss protection

Monitor positions:

• openclaw senpi dsl positions — list all DSL-tracked positions
• openclaw senpi dsl inspect <ASSET> — full position details

Runtime Setup

Step 1: Set your strategy wallet address in runtime.yaml:

sed -i 's/${WALLET_ADDRESS}/<STRATEGY_WALLET_ADDRESS>/' /data/workspace/skills/roach-strategy/runtime.yaml

Replace <STRATEGY_WALLET_ADDRESS> with the actual wallet address.

Step 2: Set telegram chat ID for notifications:

sed -i 's/${TELEGRAM_CHAT_ID}/<CHAT_ID>/' /data/workspace/skills/roach-strategy/runtime.yaml

Replace <CHAT_ID> with the actual Telegram chat ID.

Step 3: Install the runtime:

openclaw senpi runtime create --path /data/workspace/skills/roach-strategy/runtime.yaml

Step 4: Verify:

openclaw senpi runtime list

Bootstrap Gate

On EVERY session start, check config/bootstrap-complete.json. If missing:

1. Read the senpi-trading-runtime skill: cat /data/workspace/skills/senpi-trading-runtime/SKILL.md — this provides all CLI commands for runtime management and DSL position inspection.
2. Verify Senpi MCP
3. Set wallet in runtime.yaml: sed -i 's/${WALLET_ADDRESS}/ACTUAL_ADDRESS/' /data/workspace/skills/roach-strategy/runtime.yaml
4. Set Telegram in runtime.yaml: sed -i 's/${TELEGRAM_CHAT_ID}/CHAT_ID/' /data/workspace/skills/roach-strategy/runtime.yaml
5. Install runtime: openclaw senpi runtime create --path /data/workspace/skills/roach-strategy/runtime.yaml
6. Verify runtime installed: openclaw senpi runtime list
7. Remove old DSL cron (if upgrading): run openclaw crons list, delete any cron containing dsl-v5.py via openclaw crons delete <id>
8. Create scanner cron (90s, main)
9. Write config/bootstrap-complete.json
10. Send: "🪳 ROACH v1.0 online. Striker only. Stalker disabled. Waiting for explosions. Silence = no explosion."

If bootstrap exists, still verify runtime and scanner cron on every session start.

Expected Behavior

SILENCE IS CORRECT. If ROACH goes 24-48 hours without a trade, that means there were no FIRST_JUMP explosions worth taking. That's the experiment working, not a bug.

Risk Management

Notification Policy

ONLY alert: Position OPENED (Striker signal with score + reasons), position CLOSED (P&L + reason), critical error.

NEVER alert: Scanner found no Striker signals (this is normal and expected), any reasoning about whether to enable Stalker, any analysis of market conditions.

Files

License

MIT — Built by Senpi (https://senpi.ai). 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.