@substrate-ai/core 0.20.95 → 0.20.97

package/dist/persistence/schema.js

@@ -1,736 +1,71 @@
 /**
- * Consolidated schema DDL for all persistence tables.
+ * Composition-root for persistence schema initialization.
  *
- * Replaces the 11 separate SQLite migration files with a single
- * async function that creates all tables via DatabaseAdapter.
+ * `initSchema(adapter)` calls the per-subsystem `initXxxSchema` functions in
+ * the correct order (tables before views; dependencies before dependents).
  *
- * All tables use CREATE TABLE IF NOT EXISTS for idempotency.
- * Indexes and views use IF NOT EXISTS.
+ * Subsystem ownership (Ship 5, 2026-05 — per-subsystem schema split):
+ *   - core-schema.ts        — sessions, tasks, plans, execution log, cost entries, signals, schema_migrations + ready_tasks/session_cost_summary views
+ *   - pipeline-schema.ts    — pipeline_runs, decisions, requirements, constraints, artifacts, token_usage, run_metrics, story_metrics
+ *   - monitor-schema.ts     — task_metrics, performance_aggregates, routing_recommendations (main DB; monitor.db is separate)
+ *   - state-schema.ts       — stories, contracts, metrics, dispatch_log, build_results, review_verdicts, _schema_version (legacy)
+ *   - repo-map-schema.ts    — repo_map_symbols, repo_map_meta
+ *   - telemetry-schema.ts   — turn_analysis, efficiency_scores, recommendations, category_stats, consumer_stats
+ *   - work-graph-schema.ts  — wg_stories, story_dependencies, ready_stories view
  *
- * Schema covers:
- *   - Core: sessions, tasks, task_dependencies, execution_log
- *   - Cost: cost_entries
- *   - Pipeline: pipeline_runs, decisions, requirements, constraints, artifacts, token_usage
- *   - Plans: plans, plan_versions
- *   - Signals: session_signals
- *   - Metrics: run_metrics, story_metrics
- *   - Monitor: task_metrics, performance_aggregates, routing_recommendations
- *   - Telemetry: turn_analysis, efficiency_scores, recommendations, category_stats, consumer_stats
+ * All per-subsystem inits are idempotent (CREATE TABLE IF NOT EXISTS, INSERT
+ * IGNORE seeds, try/catch ALTER ADD COLUMN migrations).
+ *
+ * Composition root order matters for views:
+ *   - `ready_tasks` and `session_cost_summary` (initCoreViews) reference
+ *     `tasks` + `sessions` — must run AFTER initCoreSchema.
+ *   - `ready_stories` (in initWorkGraphSchema) references wg_stories +
+ *     story_dependencies, both defined inside the same function — self-contained.
  */
+import { initCoreSchema, initCoreViews } from './core-schema.js';
+import { initPipelineSchema } from './pipeline-schema.js';
+import { initMonitorSchema } from './monitor-schema.js';
+import { initStateSchema } from './state-schema.js';
+import { initRepoMapSchema } from './repo-map-schema.js';
+import { initTelemetrySchema } from './telemetry-schema.js';
+import { initWorkGraphSchema } from './work-graph-schema.js';
 /**
  * Initialize all persistence tables on the given adapter.
  * Idempotent — safe to call multiple times.
  */
 export async function initSchema(adapter) {
-    // -- Core tables (migration 001 + 003) ------------------------------------
-    await adapter.exec(`
-    CREATE TABLE IF NOT EXISTS sessions (
-      id                VARCHAR(255) PRIMARY KEY,
-      name              TEXT,
-      graph_file        TEXT NOT NULL,
-      status            VARCHAR(32) NOT NULL DEFAULT 'active',
-      budget_usd        DOUBLE,
-      total_cost_usd    DOUBLE NOT NULL DEFAULT 0.0,
-      planning_cost_usd DOUBLE NOT NULL DEFAULT 0.0,
-      config_snapshot   TEXT,
-      base_branch       TEXT NOT NULL DEFAULT 'main',
-      plan_source       TEXT,
-      planning_agent    TEXT,
-      planning_costs_count_against_budget INTEGER NOT NULL DEFAULT 0,
-      created_at        DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP,
-      updated_at        DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP
-    )
-  `);
-    await adapter.exec(`
-    CREATE TABLE IF NOT EXISTS tasks (
-      id              VARCHAR(255) PRIMARY KEY,
-      session_id      VARCHAR(255) NOT NULL,
-      name            TEXT NOT NULL,
-      description     TEXT,
-      prompt          TEXT NOT NULL,
-      status          VARCHAR(32) NOT NULL DEFAULT 'pending',
-      agent           VARCHAR(128),
-      model           TEXT,
-      billing_mode    VARCHAR(32),
-      worktree_path   TEXT,
-      worktree_branch TEXT,
-      worktree_cleaned_at TEXT,
-      worker_id       TEXT,
-      budget_usd      DOUBLE,
-      cost_usd        DOUBLE NOT NULL DEFAULT 0.0,
-      input_tokens    INTEGER NOT NULL DEFAULT 0,
-      output_tokens   INTEGER NOT NULL DEFAULT 0,
-      result          TEXT,
-      error           TEXT,
-      exit_code       INTEGER,
-      retry_count     INTEGER NOT NULL DEFAULT 0,
-      max_retries     INTEGER NOT NULL DEFAULT 2,
-      timeout_ms      INTEGER,
-      task_type       TEXT,
-      metadata        TEXT,
-      merge_status    TEXT,
-      merged_files    TEXT,
-      conflict_files  TEXT,
-      budget_exceeded INTEGER NOT NULL DEFAULT 0,
-      started_at      TEXT,
-      completed_at    TEXT,
-      created_at      DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP,
-      updated_at      DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP
-    )
-  `);
-    await adapter.exec('CREATE INDEX IF NOT EXISTS idx_tasks_session ON tasks(session_id)');
-    await adapter.exec('CREATE INDEX IF NOT EXISTS idx_tasks_status ON tasks(status)');
-    await adapter.exec('CREATE INDEX IF NOT EXISTS idx_tasks_agent ON tasks(agent)');
-    await adapter.exec('CREATE INDEX IF NOT EXISTS idx_tasks_session_status ON tasks(session_id, status)');
-    await adapter.exec(`
-    CREATE TABLE IF NOT EXISTS task_dependencies (
-      task_id    VARCHAR(255) NOT NULL,
-      depends_on VARCHAR(255) NOT NULL,
-      PRIMARY KEY (task_id, depends_on),
-      CHECK (task_id != depends_on)
-    )
-  `);
-    await adapter.exec('CREATE INDEX IF NOT EXISTS idx_deps_depends_on ON task_dependencies(depends_on)');
-    await adapter.exec(`
-    CREATE TABLE IF NOT EXISTS execution_log (
-      id         INTEGER PRIMARY KEY AUTO_INCREMENT,
-      session_id VARCHAR(255) NOT NULL,
-      task_id    VARCHAR(255),
-      event      VARCHAR(128) NOT NULL,
-      old_status VARCHAR(32),
-      new_status VARCHAR(32),
-      agent      VARCHAR(128),
-      cost_usd   DOUBLE,
-      data       TEXT,
-      timestamp  DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP
-    )
-  `);
-    await adapter.exec('CREATE INDEX IF NOT EXISTS idx_log_session ON execution_log(session_id)');
-    await adapter.exec('CREATE INDEX IF NOT EXISTS idx_log_task ON execution_log(task_id)');
-    await adapter.exec('CREATE INDEX IF NOT EXISTS idx_log_event ON execution_log(event)');
-    await adapter.exec('CREATE INDEX IF NOT EXISTS idx_log_timestamp ON execution_log(timestamp)');
-    // -- Cost entries (migration 001 + 002) -----------------------------------
-    await adapter.exec(`
-    CREATE TABLE IF NOT EXISTS cost_entries (
-      id             INTEGER PRIMARY KEY AUTO_INCREMENT,
-      session_id     VARCHAR(255) NOT NULL,
-      task_id        VARCHAR(255),
-      agent          VARCHAR(128) NOT NULL,
-      billing_mode   VARCHAR(32) NOT NULL,
-      category       VARCHAR(64) NOT NULL DEFAULT 'execution',
-      provider       VARCHAR(64) NOT NULL DEFAULT 'unknown',
-      input_tokens   INTEGER NOT NULL DEFAULT 0,
-      output_tokens  INTEGER NOT NULL DEFAULT 0,
-      estimated_cost DOUBLE NOT NULL DEFAULT 0.0,
-      actual_cost    DOUBLE,
-      savings_usd    DOUBLE NOT NULL DEFAULT 0.0,
-      model          TEXT,
-      timestamp      DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP
-    )
-  `);
-    await adapter.exec('CREATE INDEX IF NOT EXISTS idx_cost_session ON cost_entries(session_id)');
-    await adapter.exec('CREATE INDEX IF NOT EXISTS idx_cost_task ON cost_entries(task_id)');
-    await adapter.exec('CREATE INDEX IF NOT EXISTS idx_cost_category ON cost_entries(category)');
-    await adapter.exec('CREATE INDEX IF NOT EXISTS idx_cost_entries_session_task ON cost_entries(session_id, task_id)');
-    await adapter.exec('CREATE INDEX IF NOT EXISTS idx_cost_entries_provider ON cost_entries(provider)');
-    await adapter.exec('CREATE INDEX IF NOT EXISTS idx_cost_session_agent ON cost_entries(session_id, agent)');
-    await adapter.exec('CREATE INDEX IF NOT EXISTS idx_cost_agent ON cost_entries(agent)');
-    // -- Session signals (migration 004) --------------------------------------
-    await adapter.exec(`
-    CREATE TABLE IF NOT EXISTS session_signals (
-      id           INTEGER PRIMARY KEY AUTO_INCREMENT,
-      session_id   VARCHAR(255) NOT NULL,
-      \`signal\`   VARCHAR(16) NOT NULL CHECK(\`signal\` IN ('pause', 'resume', 'cancel')),
-      created_at   DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP,
-      processed_at TEXT
-    )
-  `);
-    // -- Plans (migration 005 + 006) ------------------------------------------
-    await adapter.exec(`
-    CREATE TABLE IF NOT EXISTS plans (
-      id                 VARCHAR(255) PRIMARY KEY,
-      description        TEXT NOT NULL,
-      task_count         INTEGER NOT NULL DEFAULT 0,
-      estimated_cost_usd DOUBLE NOT NULL DEFAULT 0.0,
-      planning_agent     VARCHAR(128) NOT NULL,
-      plan_yaml          TEXT NOT NULL,
-      status             VARCHAR(32) NOT NULL DEFAULT 'draft',
-      current_version    INTEGER NOT NULL DEFAULT 1,
-      created_at         DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP,
-      updated_at         DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP
-    )
-  `);
-    await adapter.exec('CREATE INDEX IF NOT EXISTS idx_plans_status ON plans(status)');
-    await adapter.exec('CREATE INDEX IF NOT EXISTS idx_plans_created ON plans(created_at)');
-    await adapter.exec(`
-    CREATE TABLE IF NOT EXISTS plan_versions (
-      plan_id           VARCHAR(255) NOT NULL,
-      version           INTEGER NOT NULL,
-      task_graph_yaml   TEXT NOT NULL,
-      feedback_used     TEXT,
-      planning_cost_usd DOUBLE NOT NULL DEFAULT 0.0,
-      created_at        DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP,
-      PRIMARY KEY (plan_id, version)
-    )
-  `);
-    await adapter.exec('CREATE INDEX IF NOT EXISTS idx_plan_versions_plan_id ON plan_versions(plan_id)');
-    // -- Pipeline runs + decisions (migration 007 + 008 final shapes) ---------
-    await adapter.exec(`
-    CREATE TABLE IF NOT EXISTS pipeline_runs (
-      id               VARCHAR(255) PRIMARY KEY,
-      methodology      VARCHAR(128) NOT NULL,
-      current_phase    VARCHAR(64),
-      status           VARCHAR(32) NOT NULL DEFAULT 'running'
-                       CHECK(status IN ('running','paused','completed','failed','stopped')),
-      config_json      TEXT,
-      token_usage_json TEXT,
-      parent_run_id    VARCHAR(255),
-      created_at       DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP,
-      updated_at       DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP
-    )
-  `);
-    await adapter.exec('CREATE INDEX IF NOT EXISTS idx_pipeline_runs_status ON pipeline_runs(status)');
-    await adapter.exec('CREATE INDEX IF NOT EXISTS idx_pipeline_runs_parent_run_id ON pipeline_runs(parent_run_id)');
-    await adapter.exec(`
-    CREATE TABLE IF NOT EXISTS decisions (
-      id              VARCHAR(255) PRIMARY KEY,
-      pipeline_run_id VARCHAR(255),
-      phase           VARCHAR(64) NOT NULL,
-      category        VARCHAR(64) NOT NULL,
-      \`key\`         VARCHAR(255) NOT NULL,
-      value           TEXT NOT NULL,
-      rationale       TEXT,
-      superseded_by   VARCHAR(255),
-      created_at      DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP,
-      updated_at      DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP
-    )
-  `);
-    await adapter.exec('CREATE INDEX IF NOT EXISTS idx_decisions_phase ON decisions(phase)');
-    await adapter.exec('CREATE INDEX IF NOT EXISTS idx_decisions_key ON decisions(phase, `key`)');
-    await adapter.exec('CREATE INDEX IF NOT EXISTS idx_decisions_superseded_by ON decisions(superseded_by)');
-    await adapter.exec(`
-    CREATE TABLE IF NOT EXISTS requirements (
-      id              VARCHAR(255) PRIMARY KEY,
-      pipeline_run_id VARCHAR(255),
-      source          VARCHAR(128) NOT NULL,
-      type            VARCHAR(32) NOT NULL CHECK(type IN ('functional','non_functional','constraint')),
-      description     TEXT NOT NULL,
-      priority        VARCHAR(16) NOT NULL CHECK(priority IN ('must','should','could','wont')),
-      status          VARCHAR(32) NOT NULL DEFAULT 'active',
-      created_at      DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP
-    )
-  `);
-    await adapter.exec('CREATE INDEX IF NOT EXISTS idx_requirements_type ON requirements(type)');
-    await adapter.exec('CREATE INDEX IF NOT EXISTS idx_requirements_status ON requirements(status)');
-    await adapter.exec(`
-    CREATE TABLE IF NOT EXISTS constraints (
-      id              VARCHAR(255) PRIMARY KEY,
-      pipeline_run_id VARCHAR(255),
-      category        VARCHAR(64) NOT NULL,
-      description     TEXT NOT NULL,
-      source          VARCHAR(128) NOT NULL,
-      created_at      DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP
-    )
-  `);
-    await adapter.exec(`
-    CREATE TABLE IF NOT EXISTS artifacts (
-      id              VARCHAR(255) PRIMARY KEY,
-      pipeline_run_id VARCHAR(255),
-      phase           VARCHAR(64) NOT NULL,
-      type            VARCHAR(128) NOT NULL,
-      path            TEXT NOT NULL,
-      content_hash    TEXT,
-      summary         TEXT,
-      created_at      DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP
-    )
-  `);
-    await adapter.exec('CREATE INDEX IF NOT EXISTS idx_artifacts_phase ON artifacts(phase)');
-    await adapter.exec(`
-    CREATE TABLE IF NOT EXISTS token_usage (
-      id              INTEGER PRIMARY KEY AUTO_INCREMENT,
-      pipeline_run_id VARCHAR(255),
-      phase           VARCHAR(64) NOT NULL,
-      agent           VARCHAR(128) NOT NULL,
-      input_tokens    INTEGER NOT NULL DEFAULT 0,
-      output_tokens   INTEGER NOT NULL DEFAULT 0,
-      cost_usd        DOUBLE NOT NULL DEFAULT 0.0,
-      metadata        TEXT,
-      created_at      DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP
-    )
-  `);
-    await adapter.exec('CREATE INDEX IF NOT EXISTS idx_token_usage_run ON token_usage(pipeline_run_id)');
-    // -- Run metrics (migration 010) ------------------------------------------
-    await adapter.exec(`
-    CREATE TABLE IF NOT EXISTS run_metrics (
-      run_id              VARCHAR(255) PRIMARY KEY,
-      methodology         VARCHAR(128) NOT NULL,
-      status              VARCHAR(32) NOT NULL DEFAULT 'running',
-      started_at          TEXT NOT NULL,
-      completed_at        TEXT,
-      wall_clock_seconds  DOUBLE DEFAULT 0,
-      total_input_tokens  INTEGER DEFAULT 0,
-      total_output_tokens INTEGER DEFAULT 0,
-      total_cost_usd      DOUBLE DEFAULT 0,
-      stories_attempted   INTEGER DEFAULT 0,
-      stories_succeeded   INTEGER DEFAULT 0,
-      stories_failed      INTEGER DEFAULT 0,
-      stories_escalated   INTEGER DEFAULT 0,
-      total_review_cycles INTEGER DEFAULT 0,
-      total_dispatches    INTEGER DEFAULT 0,
-      concurrency_setting INTEGER DEFAULT 1,
-      max_concurrent_actual INTEGER DEFAULT 1,
-      restarts            INTEGER DEFAULT 0,
-      is_baseline         INTEGER DEFAULT 0,
-      created_at          DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP
-    )
-  `);
-    await adapter.exec(`
-    CREATE TABLE IF NOT EXISTS story_metrics (
-      id                  INTEGER PRIMARY KEY AUTO_INCREMENT,
-      run_id              VARCHAR(255) NOT NULL,
-      story_key           VARCHAR(255) NOT NULL,
-      result              VARCHAR(32) NOT NULL DEFAULT 'pending',
-      phase_durations_json TEXT,
-      started_at          TEXT,
-      completed_at        TEXT,
-      wall_clock_seconds  DOUBLE DEFAULT 0,
-      input_tokens        INTEGER DEFAULT 0,
-      output_tokens       INTEGER DEFAULT 0,
-      cost_usd            DOUBLE DEFAULT 0,
-      review_cycles       INTEGER DEFAULT 0,
-      dispatches          INTEGER DEFAULT 0,
-      created_at          DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP,
-      UNIQUE(run_id, story_key)
-    )
-  `);
-    // -- story_metrics agent/model columns (mesh telemetry enrichment) ---------
-    for (const col of ['primary_agent_id VARCHAR(64)', 'primary_model VARCHAR(128)', 'dispatch_agents_json TEXT']) {
-        try {
-            await adapter.exec(`ALTER TABLE story_metrics ADD COLUMN ${col}`);
-        }
-        catch { /* column already exists */ }
-    }
-    // -- Monitor tables (from 001-monitor-schema) -----------------------------
-    await adapter.exec(`
-    CREATE TABLE IF NOT EXISTS task_metrics (
-      task_id        VARCHAR(255) NOT NULL,
-      agent          VARCHAR(128) NOT NULL,
-      task_type      VARCHAR(128) NOT NULL,
-      outcome        VARCHAR(16)  NOT NULL CHECK(outcome IN ('success', 'failure')),
-      failure_reason TEXT,
-      input_tokens   INTEGER NOT NULL DEFAULT 0,
-      output_tokens  INTEGER NOT NULL DEFAULT 0,
-      duration_ms    INTEGER NOT NULL DEFAULT 0,
-      cost           DOUBLE  NOT NULL DEFAULT 0.0,
-      estimated_cost DOUBLE  NOT NULL DEFAULT 0.0,
-      billing_mode   VARCHAR(32) NOT NULL DEFAULT 'api',
-      retries        INTEGER NOT NULL DEFAULT 0,
-      recorded_at    DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP,
-      PRIMARY KEY (task_id, recorded_at)
-    )
-  `);
-    await adapter.exec('CREATE INDEX IF NOT EXISTS idx_tm_agent ON task_metrics(agent)');
-    await adapter.exec('CREATE INDEX IF NOT EXISTS idx_tm_task_type ON task_metrics(task_type)');
-    await adapter.exec('CREATE INDEX IF NOT EXISTS idx_tm_recorded_at ON task_metrics(recorded_at)');
-    await adapter.exec('CREATE INDEX IF NOT EXISTS idx_tm_agent_type ON task_metrics(agent, task_type)');
-    await adapter.exec(`
-    CREATE TABLE IF NOT EXISTS performance_aggregates (
-      agent              VARCHAR(255) NOT NULL,
-      task_type          VARCHAR(255) NOT NULL,
-      total_tasks        INTEGER NOT NULL DEFAULT 0,
-      successful_tasks   INTEGER NOT NULL DEFAULT 0,
-      failed_tasks       INTEGER NOT NULL DEFAULT 0,
-      total_input_tokens INTEGER NOT NULL DEFAULT 0,
-      total_output_tokens INTEGER NOT NULL DEFAULT 0,
-      total_duration_ms  INTEGER NOT NULL DEFAULT 0,
-      total_cost         DOUBLE  NOT NULL DEFAULT 0.0,
-      total_retries      INTEGER NOT NULL DEFAULT 0,
-      last_updated       DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP,
-      PRIMARY KEY (agent, task_type)
-    )
-  `);
-    await adapter.exec(`
-    CREATE TABLE IF NOT EXISTS routing_recommendations (
-      id                INTEGER PRIMARY KEY AUTO_INCREMENT,
-      task_type         VARCHAR(128) NOT NULL,
-      current_agent     VARCHAR(128) NOT NULL,
-      recommended_agent VARCHAR(128) NOT NULL,
-      reason            TEXT,
-      confidence        DOUBLE  NOT NULL DEFAULT 0.0,
-      supporting_data   TEXT,
-      generated_at      DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP,
-      expires_at        TEXT
-    )
-  `);
-    // -- Telemetry tables (migration 011) -------------------------------------
-    await adapter.exec(`
-    CREATE TABLE IF NOT EXISTS turn_analysis (
-      story_key         VARCHAR(64)    NOT NULL,
-      span_id           VARCHAR(128)   NOT NULL,
-      turn_number       INTEGER        NOT NULL,
-      name              VARCHAR(255)   NOT NULL DEFAULT '',
-      timestamp         BIGINT         NOT NULL DEFAULT 0,
-      source            VARCHAR(32)    NOT NULL DEFAULT '',
-      model             VARCHAR(64),
-      input_tokens      INTEGER        NOT NULL DEFAULT 0,
-      output_tokens     INTEGER        NOT NULL DEFAULT 0,
-      cache_read_tokens INTEGER        NOT NULL DEFAULT 0,
-      fresh_tokens      INTEGER        NOT NULL DEFAULT 0,
-      cache_hit_rate    DOUBLE         NOT NULL DEFAULT 0,
-      cost_usd          DOUBLE         NOT NULL DEFAULT 0,
-      duration_ms       INTEGER        NOT NULL DEFAULT 0,
-      context_size      INTEGER        NOT NULL DEFAULT 0,
-      context_delta     INTEGER        NOT NULL DEFAULT 0,
-      tool_name         VARCHAR(128),
-      is_context_spike  BOOLEAN        NOT NULL DEFAULT 0,
-      child_spans_json  TEXT           NOT NULL DEFAULT '[]',
-      task_type         VARCHAR(64),
-      phase             VARCHAR(64),
-      dispatch_id       VARCHAR(64),
-      PRIMARY KEY (story_key, span_id)
-    )
-  `);
-    await adapter.exec('CREATE INDEX IF NOT EXISTS idx_turn_analysis_story ON turn_analysis (story_key, turn_number)');
-    // Migration: add dispatch context columns for existing repos (Story 30-1)
-    for (const col of ['task_type', 'phase', 'dispatch_id']) {
-        try {
-            await adapter.exec(`ALTER TABLE turn_analysis ADD COLUMN ${col} VARCHAR(64)`);
-        }
-        catch { /* column already exists */ }
-    }
-    await adapter.exec(`
-    CREATE TABLE IF NOT EXISTS efficiency_scores (
-      story_key                     VARCHAR(64)  NOT NULL,
-      timestamp                     BIGINT       NOT NULL,
-      composite_score               INTEGER      NOT NULL DEFAULT 0,
-      cache_hit_sub_score           DOUBLE       NOT NULL DEFAULT 0,
-      io_ratio_sub_score            DOUBLE       NOT NULL DEFAULT 0,
-      context_management_sub_score  DOUBLE       NOT NULL DEFAULT 0,
-      avg_cache_hit_rate            DOUBLE       NOT NULL DEFAULT 0,
-      avg_io_ratio                  DOUBLE       NOT NULL DEFAULT 0,
-      context_spike_count           INTEGER      NOT NULL DEFAULT 0,
-      total_turns                   INTEGER      NOT NULL DEFAULT 0,
-      per_model_json                TEXT         NOT NULL DEFAULT '[]',
-      per_source_json               TEXT         NOT NULL DEFAULT '[]',
-      dispatch_id                   TEXT,
-      task_type                     TEXT,
-      phase                         TEXT,
-      PRIMARY KEY (story_key, timestamp)
-    )
-  `);
-    // Migration: add dispatch context columns for existing repos (Story 30-3)
-    for (const col of ['dispatch_id', 'task_type', 'phase']) {
-        try {
-            await adapter.exec(`ALTER TABLE efficiency_scores ADD COLUMN ${col} TEXT`);
-        }
-        catch { /* column already exists */ }
-    }
-    await adapter.exec(`
-    CREATE TABLE IF NOT EXISTS recommendations (
-      id                       VARCHAR(16)   NOT NULL,
-      story_key                VARCHAR(64)   NOT NULL,
-      sprint_id                VARCHAR(64),
-      rule_id                  VARCHAR(64)   NOT NULL,
-      severity                 VARCHAR(16)   NOT NULL,
-      title                    TEXT          NOT NULL,
-      description              TEXT          NOT NULL,
-      potential_savings_tokens INTEGER,
-      potential_savings_usd    DOUBLE,
-      action_target            TEXT,
-      generated_at             VARCHAR(32)   NOT NULL,
-      PRIMARY KEY (id)
-    )
-  `);
-    await adapter.exec(`
-    CREATE TABLE IF NOT EXISTS category_stats (
-      story_key            VARCHAR(100)   NOT NULL,
-      category             VARCHAR(30)    NOT NULL,
-      total_tokens         BIGINT         NOT NULL DEFAULT 0,
-      percentage           DECIMAL(6,3)   NOT NULL DEFAULT 0,
-      event_count          INTEGER        NOT NULL DEFAULT 0,
-      avg_tokens_per_event DECIMAL(12,2)  NOT NULL DEFAULT 0,
-      trend                VARCHAR(10)    NOT NULL DEFAULT 'stable',
-      PRIMARY KEY (story_key, category)
-    )
-  `);
-    await adapter.exec(`
-    CREATE TABLE IF NOT EXISTS consumer_stats (
-      story_key            VARCHAR(100)   NOT NULL,
-      consumer_key         VARCHAR(300)   NOT NULL,
-      category             VARCHAR(30)    NOT NULL,
-      total_tokens         BIGINT         NOT NULL DEFAULT 0,
-      percentage           DECIMAL(6,3)   NOT NULL DEFAULT 0,
-      event_count          INTEGER        NOT NULL DEFAULT 0,
-      top_invocations_json TEXT,
-      PRIMARY KEY (story_key, consumer_key)
-    )
-  `);
-    // -- Work-graph tables (Epic 31-1) ----------------------------------------
-    // Pre-v0.20.90, these lived ONLY in src/modules/state/schema.sql which is
-    // applied at `substrate init --dolt` time. That left projects whose
-    // .substrate/state/.dolt was initialized BEFORE Epic 31-1 (~2026-04)
-    // permanently missing wg_stories + story_dependencies — substrate run's
-    // initSchema didn't know about them, so the auto-migration never fired.
-    // Empirical discovery 2026-05-11: ynab (initialized 2026-03-10 at
-    // _schema_version=5) was missing all three of these despite the operator
-    // having upgraded substrate many times. Moving the DDL here makes
-    // initSchema the single contract for "what tables substrate needs."
-    await adapter.exec(`
-    CREATE TABLE IF NOT EXISTS wg_stories (
-      story_key    VARCHAR(20)   NOT NULL,
-      epic         VARCHAR(20)   NOT NULL,
-      title        VARCHAR(255),
-      status       VARCHAR(30)   NOT NULL DEFAULT 'planned',
-      spec_path    VARCHAR(500),
-      created_at   DATETIME,
-      updated_at   DATETIME,
-      completed_at DATETIME,
-      PRIMARY KEY (story_key)
-    )
-  `);
-    await adapter.exec('CREATE INDEX IF NOT EXISTS idx_wg_stories_epic ON wg_stories (epic)');
-    await adapter.exec(`
-    CREATE TABLE IF NOT EXISTS story_dependencies (
-      story_key       VARCHAR(50)   NOT NULL,
-      depends_on      VARCHAR(50)   NOT NULL,
-      dependency_type VARCHAR(50)   NOT NULL DEFAULT 'blocks',
-      source          VARCHAR(50)   NOT NULL DEFAULT 'explicit',
-      created_at      DATETIME,
-      PRIMARY KEY (story_key, depends_on)
-    )
-  `);
-    // Migration: story_dependencies.created_at was added in v0.12.0. For
-    // projects whose schema predates that, ensure the column exists. Idempotent
-    // try/catch — silently no-ops if the column is already present.
-    try {
-        await adapter.exec('ALTER TABLE story_dependencies ADD COLUMN created_at DATETIME');
-    }
-    catch { /* column already exists */ }
-    // -- Views ----------------------------------------------------------------
-    // NOTE: Views use JOINs and aggregation. They work with Dolt
-    // but NOT with InMemoryDatabaseAdapter. For InMemory backend, views are
-    // skipped silently (CREATE VIEW is an unknown statement to InMemory).
-    // ready_stories view (Epic 31-1) — stories ready to dispatch (planned/ready
-    // status AND all blocking dependencies are complete). InMemoryDatabaseAdapter
-    // explicitly no-ops CREATE VIEW (memory-adapter.ts:120-123), so this works
-    // on both backends without a try/catch wrapper — and a wrapper would hide
-    // genuine Dolt-side errors. Uses `CREATE VIEW IF NOT EXISTS` rather than
-    // `CREATE OR REPLACE` to match the existing ready_tasks pattern and avoid
-    // re-running the view definition on every initSchema call.
-    await adapter.exec(`
-    CREATE VIEW IF NOT EXISTS ready_stories AS
-      SELECT s.* FROM wg_stories s
-      WHERE s.status IN ('planned', 'ready')
-        AND NOT EXISTS (
-          SELECT 1 FROM story_dependencies d
-          JOIN wg_stories dep ON dep.story_key = d.depends_on
-          WHERE d.story_key = s.story_key
-            AND d.dependency_type = 'blocks'
-            AND dep.status <> 'complete'
-        )
-  `);
-    await adapter.exec(`
-    CREATE VIEW IF NOT EXISTS ready_tasks AS
-    SELECT t.* FROM tasks t
-    WHERE t.status = 'pending'
-      AND NOT EXISTS (
-        SELECT 1 FROM task_dependencies td
-        JOIN tasks dep ON dep.id = td.depends_on
-        WHERE td.task_id = t.id
-          AND dep.status NOT IN ('completed', 'cancelled')
-      )
-  `);
-    await adapter.exec(`
-    CREATE VIEW IF NOT EXISTS session_cost_summary AS
-    SELECT
-      s.id   AS session_id,
-      s.name AS session_name,
-      COUNT(DISTINCT t.id) AS total_tasks,
-      SUM(CASE WHEN t.status = 'completed' THEN 1 ELSE 0 END) AS completed_tasks,
-      SUM(CASE WHEN t.status = 'failed'   THEN 1 ELSE 0 END) AS failed_tasks,
-      SUM(CASE WHEN t.status = 'running'  THEN 1 ELSE 0 END) AS running_tasks,
-      COALESCE(SUM(t.cost_usd), 0) AS total_cost_usd,
-      SUM(CASE WHEN t.billing_mode = 'subscription' THEN t.cost_usd ELSE 0 END) AS subscription_cost_usd,
-      SUM(CASE WHEN t.billing_mode = 'api'          THEN t.cost_usd ELSE 0 END) AS api_cost_usd,
-      s.planning_cost_usd
-    FROM sessions s
-    LEFT JOIN tasks t ON t.session_id = s.id
-    GROUP BY s.id
-  `);
-    // -- Schema migration tracking table (for future use) --------------------
-    await adapter.exec(`
-    CREATE TABLE IF NOT EXISTS schema_migrations (
-      version    INTEGER PRIMARY KEY,
-      name       TEXT    NOT NULL,
-      applied_at DATETIME NOT NULL DEFAULT CURRENT_TIMESTAMP
-    )
-  `);
-    // -- Legacy state tables (ported from schema.sql in Ship 3, 2026-05) ------
-    // These were previously created only at `substrate init --dolt` time via
-    // schema.sql. Ship 3 moved them into initSchema so initSchema is the single
-    // contract for "what tables substrate needs," and substrate init --dolt
-    // no longer applies a separate DDL file. The Ship 2 regression gate
-    // (test/persistence/full-init-integration.test.ts) enforces that this
-    // set + the rest of initSchema produces the same table union schema.sql did.
-    //
-    // Functional status of these tables: stories/contracts/metrics/dispatch_log/
-    // build_results/review_verdicts had ZERO rows in every audited production
-    // project (ynab, quant). They survive because Ship 7 will decide their
-    // fate; deleting them outright would break operator commands that still
-    // expect them and risk an inconsistent post-init schema for existing repos.
-    await adapter.exec(`
-    CREATE TABLE IF NOT EXISTS stories (
-      story_key       VARCHAR(100)   NOT NULL,
-      sprint          VARCHAR(50),
-      status          VARCHAR(30)    NOT NULL DEFAULT 'PENDING',
-      phase           VARCHAR(30)    NOT NULL DEFAULT 'PENDING',
-      ac_results      JSON,
-      error_message   TEXT,
-      created_at      DATETIME,
-      updated_at      DATETIME,
-      completed_at    DATETIME,
-      PRIMARY KEY (story_key)
-    )
-  `);
-    await adapter.exec(`
-    CREATE TABLE IF NOT EXISTS contracts (
-      story_key    VARCHAR(100)   NOT NULL,
-      name         VARCHAR(200)   NOT NULL,
-      direction    VARCHAR(20)    NOT NULL,
-      schema_path  VARCHAR(500),
-      transport    VARCHAR(200),
-      recorded_at  DATETIME,
-      PRIMARY KEY (story_key, name, direction)
-    )
-  `);
-    await adapter.exec(`
-    CREATE TABLE IF NOT EXISTS metrics (
-      story_key          VARCHAR(100)   NOT NULL,
-      task_type          VARCHAR(100)   NOT NULL,
-      recorded_at        DATETIME       NOT NULL,
-      model              VARCHAR(100),
-      tokens_in          BIGINT         NOT NULL DEFAULT 0,
-      tokens_out         BIGINT         NOT NULL DEFAULT 0,
-      cache_read_tokens  BIGINT         NOT NULL DEFAULT 0,
-      cost_usd           DECIMAL(10,6)  NOT NULL DEFAULT 0,
-      wall_clock_ms      BIGINT         NOT NULL DEFAULT 0,
-      review_cycles      INT            NOT NULL DEFAULT 0,
-      stall_count        INT            NOT NULL DEFAULT 0,
-      result             VARCHAR(30),
-      PRIMARY KEY (story_key, task_type, recorded_at)
-    )
-  `);
-    await adapter.exec(`
-    CREATE TABLE IF NOT EXISTS dispatch_log (
-      story_key      VARCHAR(100)   NOT NULL,
-      dispatched_at  DATETIME       NOT NULL,
-      branch         VARCHAR(200),
-      worker_id      VARCHAR(100),
-      result         VARCHAR(30),
-      PRIMARY KEY (story_key, dispatched_at)
-    )
-  `);
-    await adapter.exec(`
-    CREATE TABLE IF NOT EXISTS build_results (
-      story_key    VARCHAR(100)   NOT NULL,
-      timestamp    DATETIME       NOT NULL,
-      command      VARCHAR(500),
-      exit_code    INT,
-      stdout_hash  VARCHAR(64),
-      PRIMARY KEY (story_key, timestamp)
-    )
-  `);
-    await adapter.exec(`
-    CREATE TABLE IF NOT EXISTS review_verdicts (
-      story_key     VARCHAR(100)   NOT NULL,
-      timestamp     DATETIME       NOT NULL,
-      verdict       VARCHAR(30),
-      issues_count  INT            NOT NULL DEFAULT 0,
-      notes         TEXT,
-      PRIMARY KEY (story_key, timestamp)
-    )
-  `);
-    await adapter.exec(`
-    CREATE TABLE IF NOT EXISTS _schema_version (
-      version      INT            NOT NULL,
-      applied_at   DATETIME       NOT NULL DEFAULT CURRENT_TIMESTAMP,
-      description  VARCHAR(500),
-      PRIMARY KEY (version)
-    )
-  `);
-    // _schema_version seed rows — preserved verbatim from schema.sql for
-    // backward-compat with operators inspecting the version table. Ship 7
-    // will decide whether to keep the table or delete it.
-    try {
-        await adapter.exec(`INSERT IGNORE INTO _schema_version (version, description) VALUES (1, 'Initial substrate state schema')`);
-    }
-    catch { /* InMemory adapter does not support INSERT IGNORE */ }
-    try {
-        await adapter.exec(`INSERT IGNORE INTO _schema_version (version, description) VALUES (2, 'Add turn_analysis table (Epic 27-4)')`);
-    }
-    catch { /* same */ }
-    try {
-        await adapter.exec(`INSERT IGNORE INTO _schema_version (version, description) VALUES (3, 'Add category_stats and consumer_stats tables (Epic 27-5)')`);
-    }
-    catch { /* same */ }
-    try {
-        await adapter.exec(`INSERT IGNORE INTO _schema_version (version, description) VALUES (4, 'Add recommendations table (Epic 27-7)')`);
-    }
-    catch { /* same */ }
-    try {
-        await adapter.exec(`INSERT IGNORE INTO _schema_version (version, description) VALUES (5, 'Add repo_map_symbols and repo_map_meta tables (Epic 28-2)')`);
-    }
-    catch { /* same */ }
-    try {
-        await adapter.exec(`INSERT IGNORE INTO _schema_version (version, description) VALUES (6, 'Add dependencies JSON column to repo_map_symbols (Epic 28-3)')`);
-    }
-    catch { /* same */ }
-    try {
-        await adapter.exec(`INSERT IGNORE INTO _schema_version (version, description) VALUES (7, 'Add wg_stories, story_dependencies tables and ready_stories view (Epic 31-1)')`);
-    }
-    catch { /* same */ }
-    try {
-        await adapter.exec(`INSERT IGNORE INTO _schema_version (version, description) VALUES (8, 'Add task_type, phase, dispatch_id columns to turn_analysis (Story 30-1)')`);
-    }
-    catch { /* same */ }
-    try {
-        await adapter.exec(`INSERT IGNORE INTO _schema_version (version, description) VALUES (9, 'Add dispatch_id, task_type, phase columns to efficiency_scores (Story 30-3)')`);
-    }
-    catch { /* same */ }
-    // repo_map_symbols + repo_map_meta — actively used by src/modules/repo-map/storage.ts
-    await adapter.exec(`
-    CREATE TABLE IF NOT EXISTS repo_map_symbols (
-      id          BIGINT AUTO_INCREMENT NOT NULL,
-      file_path   VARCHAR(1000)         NOT NULL,
-      symbol_name VARCHAR(500)          NOT NULL,
-      symbol_kind VARCHAR(20)           NOT NULL,
-      signature   TEXT,
-      line_number INT                   NOT NULL DEFAULT 0,
-      exported    TINYINT(1)            NOT NULL DEFAULT 0,
-      file_hash   VARCHAR(64)           NOT NULL,
-      dependencies JSON,
-      PRIMARY KEY (id)
-    )
-  `);
-    await adapter.exec('CREATE INDEX IF NOT EXISTS idx_repo_map_symbols_file ON repo_map_symbols (file_path)');
-    await adapter.exec('CREATE INDEX IF NOT EXISTS idx_repo_map_symbols_kind ON repo_map_symbols (symbol_kind)');
-    await adapter.exec(`
-    CREATE TABLE IF NOT EXISTS repo_map_meta (
-      id          INT      NOT NULL DEFAULT 1,
-      commit_sha  VARCHAR(64),
-      updated_at  DATETIME,
-      file_count  INT      NOT NULL DEFAULT 0,
-      PRIMARY KEY (id)
-    )
-  `);
-    // -- Telemetry indexes (ported from schema.sql in Ship 3) ----------------
-    // These indexes were declared in schema.sql but not in the original initSchema.
-    // The Ship 2 regression gate doesn't check for individual indexes by name,
-    // but read-heavy operator commands (`substrate metrics`) depend on them
-    // for query performance against story_key columns.
-    await adapter.exec('CREATE INDEX IF NOT EXISTS idx_efficiency_story ON efficiency_scores (story_key, timestamp DESC)');
-    await adapter.exec('CREATE INDEX IF NOT EXISTS idx_recommendations_story ON recommendations (story_key, severity)');
-    await adapter.exec('CREATE INDEX IF NOT EXISTS idx_category_stats_story ON category_stats (story_key, total_tokens)');
-    await adapter.exec('CREATE INDEX IF NOT EXISTS idx_consumer_stats_story ON consumer_stats (story_key, total_tokens)');
+    // 1. Core tables (sessions, tasks, plans, ...) — must be first so the
+    //    views below can reference them.
+    await initCoreSchema(adapter);
+    // 2. Pipeline state (pipeline_runs, decisions, requirements, ...).
+    await initPipelineSchema(adapter);
+    // 3. Monitor tables on the main DB (task_metrics, performance_aggregates,
+    //    routing_recommendations). `.substrate/monitor.db` is a separate DB
+    //    managed independently by monitor-database.ts.
+    await initMonitorSchema(adapter);
+    // 4. Legacy state tables (Ship 1 excised the corresponding writes;
+    //    Ship 7 will decide their final fate). Kept for now to preserve
+    //    backward-compat with operator commands that still expect them.
+    await initStateSchema(adapter);
+    // 5. Repo-map tables (used by src/modules/repo-map/storage.ts).
+    await initRepoMapSchema(adapter);
+    // 6. Telemetry tables (turn_analysis, efficiency_scores, ...).
+    await initTelemetrySchema(adapter);
+    // 7. Work-graph tables + ready_stories view (Epic 31-1).
+    await initWorkGraphSchema(adapter);
+    // 8. Core views (ready_tasks, session_cost_summary) — must run LAST so
+    //    sessions/tasks/task_dependencies all exist.
+    await initCoreViews(adapter);
 }
+// Re-export the per-subsystem init functions so consumers can call individual
+// subsystems if they don't need the full schema. Operator CLI commands that
+// only touch one subsystem (e.g. `substrate epic-status` only needs the
+// work-graph tables) can import the narrow init function directly.
+export { initCoreSchema, initCoreViews } from './core-schema.js';
+export { initPipelineSchema } from './pipeline-schema.js';
+export { initMonitorSchema } from './monitor-schema.js';
+export { initStateSchema } from './state-schema.js';
+export { initRepoMapSchema } from './repo-map-schema.js';
+export { initTelemetrySchema } from './telemetry-schema.js';
+export { initWorkGraphSchema } from './work-graph-schema.js';
 //# sourceMappingURL=schema.js.map