devchain-cli 0.14.1 → 0.16.0

package/README.md

@@ -5,9 +5,9 @@
 [![npm version](https://img.shields.io/npm/v/devchain-cli)](https://www.npmjs.com/package/devchain-cli)
 [![License: Elastic-2.0](https://img.shields.io/badge/license-Elastic--2.0-blue)](LICENSE)
 
-**[Homepage](https://devchain.twitechlab.com/)** · **[Quick Start Guides](https://devchain.twitechlab.com/docs/quick-start-guide/)** · **[What's New in v0.12.0](https://devchain.twitechlab.com/releases/0.12.0/)**
+**[Homepage](https://devchain.cc/)** · **[Quick Start Guides](https://devchain.cc/docs/quick-start-guide/)** · **[What's New in v0.15.0](https://devchain.cc/releases/0.15.0/)**
 
-Devchain runs your AI coding agents as coordinated teams — each with their own terminal session, task queue, and chat. Group agents under team leads, assign epics, track progress on a visual board, and let teams scale themselves to match the workload. Supports Claude Code, Codex, Gemini CLI, and OpenCode out of the box.
+Devchain runs your AI coding agents as coordinated teams — each with their own terminal session, task queue, and chat. Group agents under team leads, assign epics, track progress on a visual board, and let teams scale themselves to match the workload. Supports Claude Code, Codex, OpenCode, Antigravity, and GitHub Copilot out of the box.
 
 ---
 
@@ -16,6 +16,9 @@ Devchain runs your AI coding agents as coordinated teams — each with their own
 ### Agent Teams
 Group agents into named teams with team leads that do real management. The **Builders** team grows itself with the workload and picks the right model for each task — cheaper models for routine changes, top-tier models for harder work. The **Planning** team supports parallel planning: add multiple Architects on different models and the Brainstormer gathers independent framings from each before drafting the master plan. Choose your allowed providers per team when you create the project.
 
+### Mobile App
+Drive and monitor your agent teams from your phone, in open beta on [iOS](https://testflight.apple.com/join/VSbfE1c6) and [Android](https://play.google.com/apps/testing/com.twitech.devchain.mobile). Chat with agents, answer their AskUserQuestion prompts, reassign epics and add comments on the board, watch a live terminal viewport, and get push notifications the moment a session stops or needs your input. Everything between your PC and phone is end-to-end encrypted: the cloud relay only forwards sealed data it can't read, so your transcripts and commands stay private. Smart notifications, quiet hours, and per-project forwarding keep the noise down.
+
 ### Session Reader
 Full transcript viewer for active agent sessions, built into the Chat page. See every tool call, thinking block, and response with real-time token usage, cost tracking, and compaction events. Supports Claude Code and Codex transcripts with AI turn grouping, collapsible cards, IQR-based token hotspot detection, and keyboard navigation.
 
@@ -29,7 +32,7 @@ Change any agent's provider and model on the fly from the context menu — no te
 Spin up isolated agent environments on dedicated git branches. Each worktree gets its own agent team, terminals, and chat — run multiple features in parallel and merge when ready. Worktrees run as Docker containers or local processes with full branch isolation.
 
 ### Container Isolation
-Worktree containers are provisioned automatically from the official Devchain image on [GHCR](https://github.com/orgs/TwiTech-LAB/packages/container/package/devchain). Each container has Claude Code, Codex, and Gemini CLI pre-installed, runs as a non-root user, and shares your git identity for correct commit attribution.
+Worktree containers are provisioned automatically from the official Devchain image on [GHCR](https://github.com/orgs/TwiTech-LAB/packages/container/package/devchain). Each container has Claude Code, Codex, and Antigravity CLI pre-installed, runs as a non-root user, and shares your git identity for correct commit attribution.
 
 ### Skills
 Browse and sync AI agent skills from community sources (Anthropic, OpenAI, Vercel, and more). Enable or disable skills per project and expose them to agents via MCP tools.
@@ -44,13 +47,13 @@ Real terminal streaming via tmux and WebSocket. Each agent gets its own session
 Live pre-commit diff viewer with agent integration, inline comments, `@mentions`, comment threading, and VS Code-style file navigation.
 
 ### Multi-Provider
-Works with Claude Code, OpenAI Codex, Google Gemini CLI, GLM models, and OpenCode. Switch providers per agent or switch the whole team preset with a single click.
+Works with Claude Code, OpenAI Codex, OpenCode, Antigravity, and GitHub Copilot — plus model families like GLM through provider configs. Switch providers per agent or switch the whole team preset with a single click.
 
 ### MCP Integration
 Full Model Context Protocol support. Agents get access to epics, chat, skills, reviews, and more through MCP tools — auto-configured before each session.
 
 ### Local-First
-All data stored in a local SQLite database. No cloud account required, no data leaving your machine.
+All data is stored in a local SQLite database and the platform runs entirely on your machine. No cloud account is required to use it. If you opt into the mobile app, your PC and phone connect through an end-to-end encrypted relay, so even then your plaintext data never leaves your devices.
 
 ---
 
@@ -64,8 +67,9 @@ All data stored in a local SQLite database. No cloud account required, no data l
 - **AI Provider** — at least one of:
   - [`claude`](https://claude.ai/claude-code) CLI
   - [`codex`](https://github.com/openai/codex) CLI
-  - [`gemini`](https://github.com/google-gemini/gemini-cli) CLI
   - [`opencode`](https://github.com/opencode-ai/opencode) CLI
+  - [`agy`](https://antigravity.google) CLI (Google Antigravity)
+  - [`copilot`](https://github.com/github/copilot-cli) CLI (GitHub Copilot)
 
 ---
 
@@ -96,12 +100,11 @@ devchain start --host 0.0.0.0
 devchain stop
 ```
 
-On first run, import a template from the project page to provision your agent team. Three templates are included:
+On first run, import a template from the project page to provision your agent team. Two templates are included:
 
 | Template | Agents | Best for |
 |----------|--------|----------|
 | `teams-dev` **(recommended)** | Planning team (Brainstormer + Architect), Builders team (Epic Manager + Coders), Code Reviewer | New projects — auto-scaling Builders, parallel planning, tier-aware routing |
-| `5-agents-dev` | Brainstormer, Epic Manager, SubBSM, Coder, Code Reviewer | The classic flat workflow (still bundled and supported) |
 | `3-agents-dev` | Brainstormer, SubBSM, Coder | Faster iteration with lower token overhead |
 
 ---

package/dist/cli.js

@@ -348,12 +348,14 @@ function detectInstalledProviders() {
   const detected = new Map();
   const codexPath = isBinaryInstalled('codex');
   const claudePath = isBinaryInstalled('claude');
-  const geminiPath = isBinaryInstalled('gemini');
   const opencodePath = isBinaryInstalled('opencode');
+  const agyPath = isBinaryInstalled('agy');
+  const copilotPath = isBinaryInstalled('copilot');
   if (codexPath) detected.set('codex', codexPath);
   if (claudePath) detected.set('claude', claudePath);
-  if (geminiPath) detected.set('gemini', geminiPath);
   if (opencodePath) detected.set('opencode', opencodePath);
+  if (agyPath) detected.set('agy', agyPath);
+  if (copilotPath) detected.set('copilot', copilotPath);
   return detected; // Map<name, absolutePath>
 }
 
@@ -1214,17 +1216,23 @@ async function runHostPreflightChecks(
     if (providersDetected.size === 0) {
       const guide = [
         'No provider binaries detected on PATH. Install at least one provider and retry.',
-        'Checked: "which codex", "which claude", "which gemini", and "which opencode"',
+        'Checked: "which codex", "which claude", "which opencode", "which agy", and "which copilot"',
         'Examples:',
         '  - Install Codex:    npm i -g @openai/codex (example) or follow provider docs',
         '  - Install Claude:   npm i -g @anthropic-ai/claude-code (example) or follow provider docs',
-        '  - Install Gemini:   npm i -g @google/gemini-cli (example) or follow provider docs',
+        '  - Install agy:      curl -fsSL https://antigravity.google/cli/install.sh | bash or follow provider docs',
         '  - Install OpenCode: go install github.com/opencode-ai/opencode@latest or follow provider docs',
         'Advanced: bypass with DEVCHAIN_SKIP_PROVIDER_CHECK=1',
       ].join('\n');
       if (opts.foreground) {
         log('error', 'No providers found; aborting startup', {
-          checked: ['which codex', 'which claude', 'which gemini', 'which opencode'],
+          checked: [
+            'which codex',
+            'which claude',
+            'which opencode',
+            'which agy',
+            'which copilot',
+          ],
         });
       } else {
         cli.stepDone('✗ none found');
@@ -1292,10 +1300,6 @@ function getPreferredDevApiPort(optsPort, containerMode, env = process.env) {
   return 3000;
 }
 
-function isTruthyEnvValue(value) {
-  return ['1', 'true', 'yes', 'on'].includes(String(value ?? '').trim().toLowerCase());
-}
-
 function getDevModeSpawnConfig({ containerMode, port, env = process.env }) {
   const ui = getDevUiConfig(containerMode);
   return {
@@ -1347,8 +1351,7 @@ async function main(argv) {
       'Respects LOG_LEVEL env var if set.'
     )
     .option('--dev', 'Development mode with hot reload (spawns nest --watch + vite)')
-    .option('--cloud', 'Enable Cloud and Notifications UI features for this run')
-    .option('--cloude', '[alias] Enable Cloud and Notifications UI features for this run')
+    .option('--no-cloud', 'Disable Cloud and Notifications UI features for this run')
     .option('--internal-detached-child', '[internal] Marker for detached child process')
     .action(async (rawArgs, opts) => {
       const { HostResolver } = await import(resolveSharedModuleSpecifier());
@@ -1368,11 +1371,10 @@ async function main(argv) {
         process.exit(1);
       }
       const worktreeRuntimeMode = isWorktreeRuntimeModeEnabled(worktreeRuntimeType);
-      const cloudUiEnabled = Boolean(
-        opts.cloud || opts.cloude || isTruthyEnvValue(process.env.DEVCHAIN_CLOUD_UI_ENABLED),
-      );
-      if (cloudUiEnabled) {
-        process.env.DEVCHAIN_CLOUD_UI_ENABLED = '1';
+      // Cloud UI is on by default. Commander defaults opts.cloud to `true`
+      // and sets it to `false` only when --no-cloud is passed.
+      if (opts.cloud === false) {
+        process.env.DEVCHAIN_CLOUD_UI_ENABLED = '0';
       }
 
       // Check if "help" was passed as an argument
@@ -2026,5 +2028,6 @@ module.exports = {
     normalizeWorktreeRuntimeType,
     isWorktreeRuntimeModeEnabled,
     detectGlobalPackageManager,
+    detectInstalledProviders,
   },
 };

package/dist/drizzle/0065_next_lady_bullseye.sql

@@ -0,0 +1,11 @@
+CREATE TABLE `provider_env_scopes` (
+	`provider_id` text NOT NULL,
+	`env_key` text NOT NULL,
+	`project_id` text NOT NULL,
+	`created_at` text NOT NULL,
+	PRIMARY KEY(`provider_id`, `env_key`, `project_id`),
+	FOREIGN KEY (`provider_id`) REFERENCES `providers`(`id`) ON UPDATE no action ON DELETE cascade,
+	FOREIGN KEY (`project_id`) REFERENCES `projects`(`id`) ON UPDATE no action ON DELETE cascade
+);
+--> statement-breakpoint
+CREATE INDEX `provider_env_scopes_project_id_idx` ON `provider_env_scopes` (`project_id`);