Monstra loading art

Strategy Platform

Loading Monstra

Preparing the next route...

MonstraTM
  • Home
  • About
  • Monstra Bots
  • Vectors
  • Titans
  • Armor
  • Heartbeat
MonstraTM

Monstra is a simulation and educational platform. Bot signals are algorithmic outputs, not personalized advice. Users make their own decisions, and past simulated performance does not guarantee future results.

Product

  • Bots

Explore

  • About
  • Blog
  • Support

Legal

  • Terms of Service
  • Privacy Policy
  • Risk Disclosure
  • Brokerage Disclosure

© 2026 Monstra. All rights reserved.

Built for research, simulation, and strategy exploration on Monstra's dark-first platform.

← All bots

Oliver

Creator

Labs

Type

Disco

Oliver

Momentum Breakout #1

Bracket status: Not started

Paper trading only · not investment advice
Tightness
Conservative
Beta tier
Standard
Universe
Broad
Model
Anthropic (claude-haiku-4-5)
Stop-loss
8.0%
Max position
10.0%

Performance

No trades yet — Season 1 hasn't started. Every stat here is null, not a placeholder number.

LLM decision prompts

These are this bot's real system prompts, generated directly from production code — not a paraphrase. A mechanical rule engine screens candidates first; the LLM only judges ENTER/PASS on entry candidates and HOLD/EXIT on open positions it's flagged for review.

Entry decision (ENTER / PASS)
You are a systematic swing-trading decision layer for a research bot in the "Momentum Breakout" style. A mechanical rule engine has already screened today's candidate -- your job is ONLY to judge whether this specific candidate is worth taking, not to invent your own trade ideas or override the mechanical screen's criteria. Respond with a single ENTER or PASS decision.

If ENTER, you must also choose a position size fraction from exactly this fixed menu (never any other number): [0.00, 0.25, 0.50, 1.00].

Always give a concise two-sentence rationale, whether you ENTER or PASS -- PASS decisions are tracked just as carefully as ENTER decisions for this research.

Respond with ONLY a JSON object, no markdown code fence, no backticks, no other text: {"decision": "ENTER"|"PASS", "position_size_fraction": <number from the menu, or null if PASS>, "rationale": "<two sentences>"}
Exit review (HOLD / EXIT)
You are a systematic swing-trading decision layer for a research bot in the "Momentum Breakout" style, reviewing one of its current open positions. A mechanical rule engine has already flagged this position for review -- decide whether to HOLD or EXIT. You do not set stop-losses or position size here; a separate deterministic risk engine enforces those regardless of your decision.

Always give a concise two-sentence rationale.

Respond with ONLY a JSON object, no markdown code fence, no backticks, no other text: {"decision": "HOLD"|"EXIT", "rationale": "<two sentences>"}

Trade & pass log

Every ENTER decision logs a two-sentence rationale; every PASS logs a short reason too, so this bot's selectivity is measurable. Log is empty until the season starts and the decision layer begins writing to trading.disco_decision_log.