@in-the-loop-labs/pair-review 3.9.1 → 4.0.0

package/README.md

@@ -65,7 +65,7 @@ pair-review is a local web application for keeping humans in the loop with AI co
 - **Local-First**: All data and processing happens on your machine - no cloud dependencies
 - **GitHub-Familiar UI**: Interface feels instantly familiar to GitHub users
 - **Human-in-the-Loop**: AI suggests, you decide
-- **Multiple AI Providers**: Support for Claude, Gemini, Codex, Copilot, OpenCode, Cursor, and Pi. Use your existing subscription!
+- **Multiple AI Providers**: Support for Claude, Antigravity, Codex, Copilot, OpenCode, Cursor, and Pi. Use your existing subscription!
 - **Progressive**: Start simple with manual review, add AI analysis when you need it
 
 ## Workflows
@@ -437,7 +437,7 @@ pair-review supports several environment variables for customizing behavior:
 |----------|-------------|---------|
 | `GITHUB_TOKEN` | GitHub Personal Access Token (takes precedence over config file) | - |
 | `PAIR_REVIEW_CLAUDE_CMD` | Custom command to invoke Claude CLI | `claude` |
-| `PAIR_REVIEW_GEMINI_CMD` | Custom command to invoke Gemini CLI | `gemini` |
+| `PAIR_REVIEW_ANTIGRAVITY_CMD` | Custom command to invoke Antigravity CLI | `agy` |
 | `PAIR_REVIEW_CODEX_CMD` | Custom command to invoke Codex CLI | `codex` |
 | `PAIR_REVIEW_COPILOT_CMD` | Custom command to invoke Copilot CLI | `copilot` |
 | `PAIR_REVIEW_OPENCODE_CMD` | Custom command to invoke OpenCode CLI | `opencode` |
@@ -466,8 +466,8 @@ PAIR_REVIEW_CLAUDE_CMD="/usr/local/bin/claude" pair-review 123
 # Use a wrapper command (supports multi-word commands)
 PAIR_REVIEW_CLAUDE_CMD="devx claude" pair-review 123
 
-# Use Gemini with npx
-PAIR_REVIEW_GEMINI_CMD="npx gemini" pair-review --local
+# Use a custom path for the Antigravity CLI
+PAIR_REVIEW_ANTIGRAVITY_CMD="/usr/local/bin/agy" pair-review --local
 
 # Force a specific model for this review
 PAIR_REVIEW_MODEL="opus" pair-review 123
@@ -491,7 +491,7 @@ Create a Personal Access Token (PAT) with these scopes:
 pair-review integrates with AI providers via their CLI tools:
 
 - **Claude**: Uses Claude Code CLI
-- **Gemini**: Uses Gemini CLI
+- **Antigravity**: Uses the Antigravity CLI (`agy`), the successor to the Gemini CLI. Install it with `curl -fsSL https://antigravity.google/cli/install.sh | bash` (macOS/Linux; it is not an npm package) — see the [Antigravity docs](https://antigravity.google/docs). Antigravity is an analysis-only provider: it runs as an agentic reviewer in non-interactive print mode and has **no chat/ACP mode**.
 - **Codex**: Uses Codex CLI
 - **GitHub Copilot**: Uses Copilot CLI
 - **OpenCode**: Uses OpenCode CLI (requires model configuration)
@@ -502,7 +502,7 @@ You can select your preferred provider and model in the repository settings UI.
 
 #### Built-in vs. Configurable Providers
 
-Most providers (Claude, Gemini, Codex, Copilot) come with built-in model definitions. **OpenCode and Pi are different** - they have no built-in models and require you to configure which models to use.
+Most providers (Claude, Antigravity, Codex, Copilot) come with built-in model definitions. **OpenCode and Pi are different** - they have no built-in models and require you to configure which models to use.
 
 #### Configuring Custom Models
 
@@ -648,7 +648,7 @@ The key difference from Review Council is consolidation order: Advanced consolid
 
 - **Large changesets with Opus 4.6**: Use review council instructions to tell the model to use a Team — it will spawn sub-agents to divide the work
 - **Multi-model parallel review**: Give each model a specific focus area (security, performance, correctness) via per-model instructions and get combined results
-- **Cross-model perspectives**: Run the same analysis across different providers (e.g., Claude, Gemini, GPT) — where models agree signals high-confidence findings, but unique outliers from a single model can be equally valuable, catching issue types that only that model excels at spotting
+- **Cross-model perspectives**: Run the same analysis across different providers (e.g., Claude, Antigravity, GPT) — where models agree signals high-confidence findings, but unique outliers from a single model can be equally valuable, catching issue types that only that model excels at spotting
 
 ### Chat
 
@@ -683,7 +683,7 @@ Configure your preferred models in `providers.pi.models` — see [AI Provider Co
 }
 ```
 
-Available chat provider IDs: `pi`, `claude`, `codex`, `copilot-acp`, `gemini-acp`, `opencode-acp`, `cursor-acp`. Each supports `command`, `args` (replaces defaults), `extra_args` (appends), and `env` overrides. Codex chat also supports `sandbox`: use `workspace-write` by default, or `read-only` for discussion-only sessions.
+Available chat provider IDs: `pi`, `claude`, `codex`, `copilot-acp`, `opencode-acp`, `cursor-acp`. Each supports `command`, `args` (replaces defaults), `extra_args` (appends), and `env` overrides. Codex chat also supports `sandbox`: use `workspace-write` by default, or `read-only` for discussion-only sessions. (Antigravity is an analysis-only provider — it has no ACP mode, so there is no Antigravity chat provider.)
 
 **Keyboard shortcut:** Press `p` then `c` to toggle the chat panel.
 
@@ -959,8 +959,8 @@ These commands update your MCP configuration in `~/.claude/settings.json` (user-
 
 ### Prerequisites
 
-- Node.js 20.0.0 or higher
-- [pnpm](https://pnpm.io/) 10.x
+- Node.js 22.0.0 or higher
+- [pnpm](https://pnpm.io/) 11.x
 - Git
 
 ### Running Locally

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@in-the-loop-labs/pair-review",
-  "version": "3.9.1",
+  "version": "4.0.0",
   "description": "Your AI-powered code review partner - Close the feedback loop with AI coding agents",
   "main": "src/server.js",
   "bin": {
@@ -19,8 +19,9 @@
     "LICENSE"
   ],
   "engines": {
-    "node": ">=20.0.0"
+    "node": ">=22.0.0"
   },
+  "packageManager": "pnpm@11.9.0",
   "scripts": {
     "start": "node src/server.js",
     "dev": "node bin/pair-review.js",
@@ -47,7 +48,7 @@
     "cursor",
     "copilot",
     "codex",
-    "gemini"
+    "antigravity"
   ],
   "author": "Tim Perkins",
   "license": "Apache-2.0",
@@ -63,7 +64,7 @@
     "@agentclientprotocol/sdk": "^0.14.1",
     "@modelcontextprotocol/sdk": "^1.25.3",
     "@octokit/rest": "^19.0.11",
-    "better-sqlite3": "^11.8.1",
+    "better-sqlite3": "^12.11.1",
     "express": "^4.18.2",
     "glob": "^13.0.6",
     "markdown-it": "^13.0.2",
@@ -83,10 +84,5 @@
     "jsdom": "^29.0.1",
     "supertest": "^7.1.4",
     "vitest": "^4.0.16"
-  },
-  "pnpm": {
-    "onlyBuiltDependencies": [
-      "better-sqlite3"
-    ]
   }
 }

package/plugin/.claude-plugin/plugin.json

@@ -1,6 +1,6 @@
 {
   "name": "pair-review",
-  "version": "3.9.1",
+  "version": "4.0.0",
   "description": "pair-review app integration — Open PRs and local changes in the pair-review web UI, run server-side AI analysis, and address review feedback. Requires the pair-review MCP server.",
   "author": {
     "name": "in-the-loop-labs",

package/plugin-code-critic/.claude-plugin/plugin.json

@@ -1,6 +1,6 @@
 {
   "name": "code-critic",
-  "version": "3.9.1",
+  "version": "4.0.0",
   "description": "AI-powered code review analysis — Run three-level AI analysis and implement-review-fix loops directly in your coding agent. Works standalone, no server required.",
   "author": {
     "name": "in-the-loop-labs",

package/public/js/components/AdvancedConfigTab.js

@@ -170,7 +170,7 @@ class AdvancedConfigTab {
   /**
    * Set default orchestration provider/model for new councils.
    * Falls back to 'claude'/'sonnet' if not provided.
-   * @param {string} provider - Default provider ID (e.g., 'claude', 'gemini')
+   * @param {string} provider - Default provider ID (e.g., 'claude', 'antigravity')
    * @param {string} model - Default model ID (e.g., 'sonnet', 'opus')
    */
   setDefaultOrchestration(provider, model) {

package/public/js/components/AnalysisConfigModal.js

@@ -627,7 +627,7 @@ class AnalysisConfigModal {
   selectModel(modelId) {
     // Guard: never store a model that doesn't belong to the current provider.
     // Callers can pass a model resolved from a different scope (e.g. 'opus'
-    // while the provider is 'gemini'); fall back to the provider's default so
+    // while the provider is 'antigravity'); fall back to the provider's default so
     // the UI always shows a selected card and submissions carry a valid pair.
     if (modelId && this.models.length && !this.models.some(m => m.id === modelId)) {
       const fallback = this.models.find(m => m.default) || this.models[0];
@@ -952,7 +952,7 @@ class AnalysisConfigModal {
     }
     // Only honour currentModel when it belongs to the selected provider.
     // selectProvider() has already chosen a valid model for the provider (via
-    // tier matching); applying a foreign model here (e.g. 'opus' under 'gemini')
+    // tier matching); applying a foreign model here (e.g. 'opus' under 'antigravity')
     // would leave no model card selected and submit an invalid pair.
     if (options.currentModel && this.models.some(m => m.id === options.currentModel)) {
       this.selectModel(options.currentModel);

package/public/js/local.js

@@ -346,7 +346,7 @@ class LocalManager {
         const lastCouncilId = reviewSettings.last_council_id;
 
         // Resolve provider and model as a MATCHED pair so the council/advanced tabs
-        // are never seeded with a cross-provider model (e.g. gemini + opus), which
+        // are never seeded with a cross-provider model (e.g. antigravity + opus), which
         // would blank the model <select> and be rejected by the backend.
         const providersInfo = await manager._getProvidersInfo();
         const { provider: currentProvider, model: currentModel } = window.resolveProviderModelPair([

package/public/js/pr.js

@@ -704,7 +704,7 @@ class PRManager {
     // Resolve provider and model as a MATCHED pair. Resolving each half
     // independently (repo || app || hardcoded) can mix a provider from one
     // scope with a model from another, yielding an invalid pair (e.g.
-    // gemini/opus) that startAnalysis would forward to the backend as-is.
+    // antigravity/opus) that startAnalysis would forward to the backend as-is.
     const providers = providersInfo || await this._getProvidersInfo();
     const { provider, model } = window.resolveProviderModelPair([
       { provider: repoSettings?.default_provider, model: repoSettings?.default_model },
@@ -3053,7 +3053,7 @@ class PRManager {
       ]);
 
       // Resolve provider and model as a MATCHED pair so the council/advanced tabs
-      // are never seeded with a cross-provider model (e.g. gemini + opus), which
+      // are never seeded with a cross-provider model (e.g. antigravity + opus), which
       // would blank the model <select> and be rejected by the backend.
       const providersInfo = await this._getProvidersInfo();
       const { provider: currentProvider, model: currentModel } = window.resolveProviderModelPair([
@@ -6846,7 +6846,7 @@ class PRManager {
       const lastCouncilId = reviewSettings.last_council_id;
 
       // Resolve provider and model as a MATCHED pair so the council/advanced tabs
-      // are never seeded with a cross-provider model (e.g. gemini + opus), which
+      // are never seeded with a cross-provider model (e.g. antigravity + opus), which
       // would blank the model <select> and be rejected by the backend.
       const providersInfo = await this._getProvidersInfo();
       const { provider: currentProvider, model: currentModel } = window.resolveProviderModelPair([

package/public/js/utils/provider-model.js

@@ -6,7 +6,7 @@
  * settings, then app config) and historically each half was resolved
  * independently with `repo || app || hardcoded`. When a scope overrides only
  * ONE half, the halves can come from different scopes and produce an invalid
- * pair such as `gemini` + `opus` (an Anthropic model). The modal then shows no
+ * pair such as `antigravity` + `opus` (an Anthropic model). The modal then shows no
  * selected model, and the non-modal auto-analyze path posts the broken pair
  * straight to the backend.
  *

package/src/ai/analyzer.js

@@ -342,8 +342,8 @@ A suggestion is a duplicate if it targets the same file and overlapping lines an
 class Analyzer {
   /**
    * @param {Object} database - Database instance
-   * @param {string} model - Model to use (e.g., 'opus', 'gemini-2.5-pro')
-   * @param {string} provider - Provider ID (e.g., 'claude', 'gemini'). Defaults to 'claude'.
+   * @param {string} model - Model to use (e.g., 'opus', 'gemini-3.1-pro-low')
+   * @param {string} provider - Provider ID (e.g., 'claude', 'antigravity'). Defaults to 'claude'.
    * @param {Object} providerOverrides - Per-call config overrides passed to createProvider (optional)
    * @param {Object|null} providerOverridesMap - Per-provider overrides map for council mode (provider ID → overrides)
    */
@@ -860,8 +860,8 @@ class Analyzer {
    *
    * Uses the bare command name since BIN_DIR is added to PATH in all
    * providers.  Using the absolute path causes issues with AI providers
-   * that pattern-match allowed tools (e.g., Gemini only matches
-   * 'git-diff-lines', not the full path).
+   * that pattern-match allowed shell commands by their bare command name
+   * (matching 'git-diff-lines', not the full path).
    *
    * @param {string|null} worktreePath - Path to the git worktree
    * @returns {string} e.g. `git-diff-lines` or `git-diff-lines --cwd "/path"`