abel-invest

What problem does it solve?

Abel Invest helps users find and validate high-quality trading strategies without relying on ad-hoc hand-designed ideas, by running an end-to-end workspace-first alpha discovery loop with legality checks, evidence tracking, and performance gate discipline.

Core Features & Use Cases

• Workspace-first strategy discovery: resolves or bootstraps the correct research workspace, then drives session/branch work using the workspace command prefix.
• Graph-informed candidate generation with empirical construction: uses live graph discovery when available to expand the candidate universe, then constructs and tests candidate families using bounded, data-led search rather than single-rule guessing.
• Safety, readiness, and validation orchestration: runs doctor checks (runtime/auth/edge contract readiness), prepares inputs, runs semantic preflight (debug), and records evidence in required audit surfaces like exploration_path.md and ledgers.
• Performance-targeted optimization posture: when a hard metric target is present (e.g., Sharpe/MaxDD/PnL), it switches to guarded optimization/reportability rules with explicit selection-width accounting.
• Alpha workspace continuity: supports continuing existing discovery work, sessions, branches, and rounds, including “grandma mode” conservative screening when requested.

Quick Start

Ask Abel Invest to find alpha for a specific target (for example: Give me an Abel strategy discovery for TSLA with a strong tradable goal and controlled drawdown).