docs: polish structured learnings README for issue #4

migrations/20260421_create_learnings_table.sql

@@ -0,0 +1,29 @@
+CREATE TABLE learnings (
+    guid UUID PRIMARY KEY DEFAULT gen_random_uuid(),
+    summary TEXT NOT NULL,
+    details TEXT,
+    category TEXT,
+    area TEXT,
+    status TEXT NOT NULL DEFAULT 'provisional',
+    priority TEXT DEFAULT 'medium',
+    confidence TEXT DEFAULT 'hypothesis',
+    action_required BOOLEAN DEFAULT false,
+    source_type TEXT,
+    source_ref TEXT,
+    project_id UUID REFERENCES projects(guid) ON DELETE SET NULL,
+    related_thought_id UUID REFERENCES thoughts(guid) ON DELETE SET NULL,
+    related_skill_id UUID REFERENCES skills(guid) ON DELETE SET NULL,
+    reviewed_by TEXT,
+    reviewed_at TIMESTAMP WITH TIME ZONE,
+    duplicate_of UUID REFERENCES learnings(guid) ON DELETE SET NULL,
+    supersedes UUID REFERENCES learnings(guid) ON DELETE SET NULL,
+    tags TEXT[],
+    created_at TIMESTAMP WITH TIME ZONE DEFAULT now(),
+    updated_at TIMESTAMP WITH TIME ZONE DEFAULT now()
+);
+
+CREATE INDEX idx_learnings_project_id ON learnings(project_id);
+CREATE INDEX idx_learnings_category ON learnings(category);
+CREATE INDEX idx_learnings_status ON learnings(status);
+CREATE INDEX idx_learnings_tags ON learnings USING GIN(tags);
+CREATE INDEX idx_learnings_search ON learnings USING GIN(to_tsvector('simple', summary || ' ' || coalesce(details, '')));