Edge Candidate Agent

Turns market observations and anomalies into structured research tickets that feed directly into trade-strategy-pipeline Phase I backtests. You run auto_detect_candidates.py on EOD OHLCV data, it generates prioritized tickets as YAML, then export_candidate.py converts them into strategy.yaml and metadata.json that match the edge-finder-candidate/v1 interface contract. The validate_candidate.py script checks schema compliance before handoff. Sits at the end of a four-stage workflow after hint extraction, concept synthesis, and strategy design. Useful if you're running systematic edge research and need the format boundaries enforced before expensive backtesting. The guardrails are strict: only pivot_breakout and gap_up_continuation families export, everything else stays research-only.