Brad Warren@CajunMonster

CAJNMNSTR is an evidence-governed agent for single-leg SPY options in a dedicated Alpaca PAPER account. It normalizes authenticated SIP equity data, OPRA option snapshots, account state, and event context into an Evidence Snapshot. Deterministic code calculates market features, validates freshness and provenance, and opens an Evidence Passport. OpenAI GPT-5.6 Terra sees that evidence—not broker tools—and returns LONG_CALL, LONG_PUT, or NO_TRADE with a thesis, counterargument, uncertainty, citations, and invalidation. A deterministic Referee independently returns APPROVE, REDUCE, ABSTAIN, or BLOCK. Only permitted proposals reach deterministic contract, quantity, premium, health, and session-risk gates. Across the competition window, CAJNMNSTR evaluated 101 decision epochs, produced 13 actionable candidates, and autonomously completed nine entries and nine exits. Every lifecycle reached CLOSED_BROKER_FLAT; no manual order was placed and no submission entered SUBMIT_UNKNOWN. Results were three wins and six losses: -$92 Tuesday, -$150 Wednesday, and +$332 Thursday. Lifecycle P&L was +$90; Alpaca equity finished at $100,088.55, up $88.55 from the $100,000 start. Maximum drawdown was 0.259%. This is a small, PAPER-only sample with modest profitability—not evidence of a statistically proven edge. The demonstrated result is autonomous, auditable operation with deterministic risk, evidence provenance, and verified broker-flat reconciliation.`

🌎 Online Hackathon | 💻 Dates: 28 August–4 September 2026 | 🏆 $6,000 Prize Pool Build AI trading agents on Alpaca — autonomous agents and trading apps using Alpaca's Trading API, MCP server and CLI. 🌍 Fully online hackathon. Join and build from anywhere in the world.