Strategy Add

Binding sources

• docs/FINAL_SYSTEM_VISION.md — strategies sit in L2–L4 signal/scoring space; outputs must remain explainable for L5 and observability.
• AGENTS.md — signals flow through risk (L7) and paper execution; no black-box metadata.

Purpose

Use this skill when strategy work must fit the current end-to-end pipeline:

• regime-aware scoring
• WAIT explainability metadata
• non-bypassable risk and execution-quality gates
• paper position lifecycle and portfolio reporting
• shadow analytics and operator panel summaries

Use When

Use this skill when:

• introducing a new strategy module under app/strategies
• wiring a strategy into scorer, reporting, or manual-review analytics
• extending signal metadata needed by explainability or shadow summaries

Do Not Use When

Do not use this skill for:

• isolated parameter tuning in an existing strategy
• scorer-confidence experiments that do not change raw strategy outputs
• portfolio sizing or risk-only changes

Required Checks

• Keep the strategy in its own small module.
• Return a structured signal object with action, confidence, reason, and any required explainability metadata.
• Integrate cleanly with scorer inputs and current reporting paths.
• Preserve WAIT explainability and journal readability.
• Confirm new signals still flow through existing risk, quality gate, and paper-only execution order.
• Update README only if the workflow or operator-visible behavior changed.

Expected Output

• Strategy: name and thesis.
• Integration Points: scorer, reporting, journal, or panel surfaces touched.
• Safety Check: paper-only and risk-order impact.
• Validation: exact checks run.