schema_forge_core

Plan and estimate a development task using an additive, component-based point system (Fibonacci base points per item type + risk factors). Produces a decomposed estimate, an hours range, a t-shirt size, and a list of the difficulties/risks that drive the cost. Also runs calibration: comparing an estimate against the real worklogs once a task is done, and proposing tweaks to the points table. Use when the user wants to estimate, plan, size, or score a task, asks "how long will this take", "cuánto va a costar", "estimemos esto", "puntuá esta tarea", "planificación", or wants to review why a task overran its estimate.

Verify that pending changes are behavior-preserving (innocuous) — that a refactor does NOT alter observable functionality. Use after refactoring to lower cognitive complexity, resolve SonarQube issues, rename, extract/inline functions, or apply guard clauses. Analyzes the diff hunk by hunk for behavior-changing edits, flags poor function names (generic auto-extract names, placeholders), then runs the affected tests. Optionally (opt-in) places START/END extraction markers in a too-complex function or .jsx component to plan a behavior-preserving extract-method/extract-component refactor. Triggers on: "is this safe", "inocuo", "inocuidad", "refactor", "bad function names", "did I break anything", "behavior preserving", "Sonar fix", "bajar complejidad cognitiva", "check my changes", "verify diff", "marca start/end", "marcadores para extraer", "comentarios para extraer", "extract markers", "ayudame a extraer", "WebStorm extract".