nodal-agents 0.3.9 → 0.4.0

package/README.md

@@ -204,7 +204,7 @@ pnpm deps:check   # runs locally and in CI before every release
 
 ## Status
 
-**Current release:** `0.3.9` on npm `latest`. Used daily by the
+**Current release:** `0.3.10` on npm `latest`. Used daily by the
 maintainer, stable enough for personal production. Pre-1.0 — breaking
 changes are still possible between minors.
 
@@ -220,15 +220,16 @@ changes are still possible between minors.
   Airtable) and API-key (Notion, Airtable, Apify, Firecrawl, Tavily)
 - MCP catalog — Streamable HTTP *and* stdio (local subprocess) servers, API-key auth; a growing catalogue (Stripe, n8n, Supabase, Airtable, Notion…) with a "test pending" badge on entries not yet verified live, plus add *and edit* your own custom HTTP/stdio servers from the dashboard
 - Top-level workspaces — multiple isolated entities (agents/skills/connectors/jobs/memory per workspace), switch in the sidebar
-- ROOT agent — designate an orchestrator that can create skills/agents and assign them, gated by per-grant toggles + an autonomy/approval level
+- In-app ROOT chat — talk to your workspace ROOT right in the dashboard: conversation-first (pure chat never creates a job — recall is free, the agent's memory is auto-loaded), multiple conversations with searchable history, and inline "dispatched to N agents" cards when it escalates a real action into a tracked job
+- ROOT agent — your first orchestrator automatically becomes the workspace ROOT (the single top-level agent; later orchestrators slot under it). It can create *and update* skills, create agents and assign them, and create MCP servers + API connectors — each gated by per-grant toggles + an autonomy/approval level (powers start off, opt in per grant). Provisioning verifies before it writes (an MCP server is connected and its tools listed first); skill authoring is grounded in the workspace's real tools (a linter rejects skills referencing tools the agent doesn't have)
 - Office file editing — Excel in-place edit, Word/PowerPoint create, in the agent workspace (office-editing skill)
 - Multiple filesystem folders per agent (sandboxed `file_*` tools)
 - Telegram delivery (long-poll, group filters, multi-agent routing,
   delegation gracefulness) — exactly-once delivery contract: anti-spam guard
   against runaway message loops + a guard that re-prompts (then fails loud)
   rather than completing a job without ever replying
-- Approval gates for risky tools (execute-the-approved-action on resume)
-- Cron scheduling
+- Approval gates for risky tools (execute-the-approved-action on resume) — on a chat channel, the agent sends a heads-up before the job pauses so the user knows an approval is waiting
+- Cron scheduling — trigger any automation out-of-band ("Run now") + opt-in Telegram confirmation when a scheduled job succeeds
 - `nodal-agents update` — one-command upgrade + boot version notice
 - Encryption at rest for LLM keys + MCP keys
 - Embedded Postgres distribution via npm (no external DB to install)
@@ -237,9 +238,8 @@ changes are still possible between minors.
 
 - **MCP OAuth flow** → unlocks Linear, Notion remote, GitHub remote,
   Atlassian, Sentry, and the rest of the SaaS-as-MCP ecosystem.
-- **Dashboard chat for the ROOT companion** → talk to your ROOT agent in
-  the app, multi-turn, instead of via Telegram; plus a dry-run mode and a
-  test-workflow meta-tool.
+- **Dry-run mode + a test-workflow meta-tool** → preview what the ROOT would
+  do before it runs, and let it validate an automation end-to-end.
 - **pgvector binaries bundled in the npm pack** → semantic memory search
   active out-of-the-box. Today, installs without pgvector fall back to
   keyword search (which works, just less smart for cross-vocabulary

package/cli.js

@@ -11272,7 +11272,8 @@ var init_enums = __esm({
       "cron",
       "task-board",
       "slack",
-      "discord"
+      "discord",
+      "dashboard"
     ];
     JobChannelSchema = z2.enum(JOB_CHANNELS);
     JOB_STATUSES = [
@@ -12168,12 +12169,22 @@ var init_root_agent = __esm({
     RootGrantsSchema = z19.object({
       createAgent: z19.boolean(),
       createSkill: z19.boolean(),
+      updateSkill: z19.boolean(),
       assignSkill: z19.boolean(),
+      createMcp: z19.boolean(),
+      createConnector: z19.boolean(),
       autonomy: AutonomyLevelSchema
     });
   }
 });
 
+// ../../packages/shared/src/connector-catalog.ts
+var init_connector_catalog = __esm({
+  "../../packages/shared/src/connector-catalog.ts"() {
+    "use strict";
+  }
+});
+
 // ../../packages/shared/src/index.ts
 var init_src2 = __esm({
   "../../packages/shared/src/index.ts"() {
@@ -12198,6 +12209,7 @@ var init_src2 = __esm({
     init_operation();
     init_providers();
     init_root_agent();
+    init_connector_catalog();
   }
 });
 
@@ -12530,7 +12542,7 @@ var init_jobs = __esm({
         ),
         check(
           "agent_jobs_channel_check",
-          sql`${table.channel} IN ('telegram','api','whatsapp','internal','cron','task-board','slack','discord')`
+          sql`${table.channel} IN ('telegram','api','whatsapp','internal','cron','task-board','slack','discord','dashboard')`
         )
       ]
     );
@@ -12975,6 +12987,10 @@ var init_schedules = __esm({
         nextRun: timestamp("next_run", { withTimezone: true }),
         lastStatus: text("last_status"),
         chatId: text("chat_id"),
+        // Opt-in: when true, a fired job carries a delivery target so the runner
+        // forces the agent to send the user a success confirmation before finishing.
+        // Default false → the cron runs silently (the user must opt in per schedule).
+        notifyOnSuccess: boolean("notify_on_success").notNull().default(false),
         createdAt: timestamp("created_at", { withTimezone: true }).defaultNow(),
         updatedAt: timestamp("updated_at", { withTimezone: true }).defaultNow()
       },
@@ -13285,6 +13301,56 @@ var init_agent_workspaces = __esm({
   }
 });
 
+// ../../packages/db/src/schema/chat-messages.ts
+var conversations, chatMessages;
+var init_chat_messages = __esm({
+  "../../packages/db/src/schema/chat-messages.ts"() {
+    "use strict";
+    init_pg_core();
+    init_drizzle_orm();
+    init_entities();
+    init_agents();
+    init_jobs();
+    conversations = pgTable(
+      "conversations",
+      {
+        id: uuid("id").primaryKey().defaultRandom(),
+        entityId: uuid("entity_id").references(() => entities.id, { onDelete: "cascade" }),
+        // The agent this conversation is with (the ROOT, today).
+        agentId: uuid("agent_id").notNull().references(() => agents.id, { onDelete: "cascade" }),
+        title: text("title").notNull().default(""),
+        createdAt: timestamp("created_at", { withTimezone: true }).defaultNow(),
+        // Bumped on each new turn — drives the recency sort in the sidebar.
+        updatedAt: timestamp("updated_at", { withTimezone: true }).defaultNow()
+      },
+      (table) => [
+        index("idx_conversations_entity_agent").on(table.entityId, table.agentId, table.updatedAt)
+      ]
+    );
+    chatMessages = pgTable(
+      "chat_messages",
+      {
+        id: uuid("id").primaryKey().defaultRandom(),
+        entityId: uuid("entity_id").references(() => entities.id, { onDelete: "cascade" }),
+        agentId: uuid("agent_id").notNull().references(() => agents.id, { onDelete: "cascade" }),
+        conversationId: uuid("conversation_id").references(() => conversations.id, {
+          onDelete: "cascade"
+        }),
+        role: text("role").notNull(),
+        content: text("content").notNull(),
+        // Set when this (assistant) turn escalated into a real action job — lets the
+        // UI render the job's dispatch/progress inline. NULL for pure conversation.
+        jobId: uuid("job_id").references(() => agentJobs.id, { onDelete: "set null" }),
+        createdAt: timestamp("created_at", { withTimezone: true }).defaultNow()
+      },
+      (table) => [
+        index("idx_chat_messages_conversation").on(table.conversationId, table.createdAt),
+        check("chat_messages_role_check", sql`${table.role} IN ('user','assistant')`)
+      ]
+    );
+  }
+});
+
 // ../../packages/db/src/schema/index.ts
 var schema_exports = {};
 __export(schema_exports, {
@@ -13326,8 +13392,10 @@ __export(schema_exports, {
   agents: () => agents,
   approvalRequests: () => approvalRequests,
   approvalRules: () => approvalRules,
+  chatMessages: () => chatMessages,
   configuratorSessions: () => configuratorSessions,
   connectors: () => connectors,
+  conversations: () => conversations,
   credentials: () => credentials,
   entities: () => entities,
   entityLlmKeys: () => entityLlmKeys,
@@ -13368,6 +13436,7 @@ var init_schema2 = __esm({
     init_auth();
     init_agent_connector_assignments();
     init_agent_workspaces();
+    init_chat_messages();
   }
 });
 
@@ -13428,7 +13497,9 @@ var init_credentials2 = __esm({
 var init_agents2 = __esm({
   "../../packages/db/src/repos/agents.ts"() {
     "use strict";
+    init_src2();
     init_agents();
+    init_entities();
   }
 });
 

package/migrations/0023_agent_schedules_notify_on_success.sql

@@ -0,0 +1,10 @@
+-- 0023_agent_schedules_notify_on_success.sql
+-- Per-schedule success confirmation (2026-05-31):
+--   ADD COLUMN agent_schedules.notify_on_success (boolean, not null, default false)
+--
+-- Opt-in flag. When true, the cron tick sets the fired job's chat_id so the
+-- runner forces the agent to deliver a success confirmation before completing.
+-- Idempotent (IF NOT EXISTS guard).
+
+ALTER TABLE agent_schedules
+  ADD COLUMN IF NOT EXISTS notify_on_success boolean NOT NULL DEFAULT false;

package/migrations/0024_agent_jobs_dashboard_channel.sql

@@ -0,0 +1,11 @@
+-- 0024_agent_jobs_dashboard_channel.sql
+-- V4 — in-app ROOT chat (2026-05-31):
+--   Allow 'dashboard' as an agent_jobs.channel value so the dashboard chat with
+--   the ROOT agent reuses the same job lifecycle + thread continuity as Telegram.
+--   Rebuild the CHECK constraint to include 'dashboard'. Idempotent.
+
+ALTER TABLE agent_jobs DROP CONSTRAINT IF EXISTS agent_jobs_channel_check;
+
+ALTER TABLE agent_jobs
+  ADD CONSTRAINT agent_jobs_channel_check
+  CHECK (channel IN ('telegram','api','whatsapp','internal','cron','task-board','slack','discord','dashboard'));

package/migrations/0025_agent_jobs_conversational.sql

@@ -0,0 +1,9 @@
+-- 0025_agent_jobs_conversational.sql
+-- V4 — in-app ROOT chat, "jobless" conversation model (2026-05-31):
+--   A chat turn that only used conversational tools (text + memory + delivery)
+--   is conversation, not work — it belongs in /chat, never in /jobs. The runner
+--   flags it at completion; the dashboard filters Runs/stats on this column.
+--   Default false → all existing + non-chat jobs remain visible as Runs.
+
+ALTER TABLE agent_jobs
+  ADD COLUMN IF NOT EXISTS conversational boolean NOT NULL DEFAULT false;

package/migrations/0026_chat_messages.sql

@@ -0,0 +1,19 @@
+-- 0026_chat_messages.sql
+-- V4 — conversation-first in-app chat (2026-05-31):
+--   A conversation is NOT a job. Pure chat turns (text + memory) are stored here
+--   and never create an agent_jobs row. Only an ACTION turn escalates to a job,
+--   whose id is linked via job_id so the UI shows progress inline.
+
+CREATE TABLE IF NOT EXISTS chat_messages (
+  id uuid PRIMARY KEY DEFAULT gen_random_uuid(),
+  entity_id uuid REFERENCES entities(id) ON DELETE CASCADE,
+  agent_id uuid NOT NULL REFERENCES agents(id) ON DELETE CASCADE,
+  thread_id text NOT NULL DEFAULT 'main',
+  role text NOT NULL CHECK (role IN ('user','assistant')),
+  content text NOT NULL,
+  job_id uuid REFERENCES agent_jobs(id) ON DELETE SET NULL,
+  created_at timestamptz DEFAULT now()
+);
+
+CREATE INDEX IF NOT EXISTS idx_chat_messages_thread
+  ON chat_messages (entity_id, agent_id, thread_id, created_at);

package/migrations/0027_drop_agent_jobs_conversational.sql

@@ -0,0 +1,6 @@
+-- 0027_drop_agent_jobs_conversational.sql
+-- Revert the short-lived "conversational" flag (migration 0025). The in-app chat
+-- is now conversation-first: pure chat lives in chat_messages and never creates
+-- an agent_jobs row at all, so flagging/hiding jobs is no longer needed.
+
+ALTER TABLE agent_jobs DROP COLUMN IF EXISTS conversational;

package/migrations/0028_conversations.sql

@@ -0,0 +1,27 @@
+-- 0028_conversations.sql
+-- V4 chat — multi-conversation backbone (2026-06-01):
+--   Group chat_messages under named `conversations` (the sidebar entries), so
+--   the chat gets multiple threads, "+ New", recency sort, and a clean history
+--   model. Replaces the single hardcoded thread_id on chat_messages.
+
+CREATE TABLE IF NOT EXISTS conversations (
+  id uuid PRIMARY KEY DEFAULT gen_random_uuid(),
+  entity_id uuid REFERENCES entities(id) ON DELETE CASCADE,
+  agent_id uuid NOT NULL REFERENCES agents(id) ON DELETE CASCADE,
+  title text NOT NULL DEFAULT '',
+  created_at timestamptz DEFAULT now(),
+  updated_at timestamptz DEFAULT now()
+);
+
+CREATE INDEX IF NOT EXISTS idx_conversations_entity_agent
+  ON conversations (entity_id, agent_id, updated_at);
+
+ALTER TABLE chat_messages
+  ADD COLUMN IF NOT EXISTS conversation_id uuid REFERENCES conversations(id) ON DELETE CASCADE;
+
+-- thread_id is superseded by conversation_id. Existing dev rows (if any) had
+-- thread_id='main' and get a NULL conversation_id (orphaned, not shown).
+ALTER TABLE chat_messages DROP COLUMN IF EXISTS thread_id;
+
+CREATE INDEX IF NOT EXISTS idx_chat_messages_conversation
+  ON chat_messages (conversation_id, created_at);

package/migrations/0029_backfill_root_agent.sql

@@ -0,0 +1,33 @@
+-- 0029_backfill_root_agent.sql
+-- Brique F — "ROOT = origin orchestrator" (2026-06-01).
+--
+-- Designate a ROOT for any entity that has at least one orchestrator but no
+-- root_agent_id yet: the earliest-created orchestrator wins. Additive only —
+-- entities that already have a ROOT (and its configured grants) are left
+-- untouched, so a manually-set ROOT and its powers are preserved.
+--
+-- Going forward, createAgentRepo auto-designates the first orchestrator of an
+-- entity as ROOT and forces subsequent ones under it. Grants start all-off
+-- (opt-in powers in Settings → ROOT agent); writing them explicitly matters
+-- because a null root_grants parses to all-on (un-gated meta-tools).
+--
+-- Idempotent: the `root_agent_id IS NULL` guard makes re-runs a no-op.
+
+UPDATE entities e
+SET
+  root_agent_id = sub.first_orch,
+  root_grants = '{"createAgent":false,"createSkill":false,"updateSkill":false,"assignSkill":false,"createMcp":false,"autonomy":"propose_confirm"}'::jsonb,
+  updated_at = now()
+FROM (
+  SELECT a.entity_id, a.id AS first_orch
+  FROM agents a
+  WHERE a.role = 'orchestrator'
+    AND NOT EXISTS (
+      SELECT 1 FROM agents a2
+      WHERE a2.entity_id = a.entity_id
+        AND a2.role = 'orchestrator'
+        AND (a2.created_at < a.created_at OR (a2.created_at = a.created_at AND a2.id < a.id))
+    )
+) AS sub
+WHERE e.id = sub.entity_id
+  AND e.root_agent_id IS NULL;