By tomplex
Web-based code review UI with AI-powered analysis for collaborative code review sessions
Analyze a code diff and classify every file by priority, phase, summary, and category. Use when generating LGTM analysis for a review session.
Synthesize per-file analysis into an overview, review strategy, and thematic file groupings. Use after file-classifier has produced per-file classifications.
Authors a narrated walkthrough of a code-review diff. Given a repo path, base branch, and output file path, inspects the diff and writes an ordered set of logical-change "stops" with titles, narratives, and artifact references. Use when building a walkthrough for an LGTM review session.
Use when the user asks to register a project for review, start a code review, open LGTM, analyze changes, or mentions LGTM. Also use when the user says they're done with a task and explicitly asks for a review.
Generate analysis for an LGTM review session — classifies every file by priority, phase, summary, and category, then produces an overview, review strategy, and thematic groupings. Use when the user asks to analyze a branch for review, or when code is mostly done and a review session is active.
Generate full review preparation for an LGTM session: classification (analyze) + narrated walkthrough. Convenience skill that chains /lgtm analyze and /lgtm walkthrough. Use when the user asks to prepare a review, or wants everything ready before opening the UI.
Refresh an existing LGTM analysis incrementally — re-classify only files whose diff has changed since the last analysis. Use after making commits on a long- lived branch when you want to keep the analysis layer fresh without redoing the full classification pass. Falls back to a full /lgtm analyze run if no prior analysis exists.
Generate a narrated walkthrough for an LGTM review session. Produces an ordered sequence of logical-change stops with titles, narratives, and artifact references. Use when the user asks to build a walkthrough, or when /lgtm prepare chains into this skill.
Uses power tools
Uses Bash, Write, or Edit tools
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
A browser-based code review UI for collaborating with Claude Code. Claude registers a project, seeds inline comments on the diff, adds documents for review, and runs analysis to prioritize files. You review everything in the browser and submit feedback that Claude reads and acts on.

The simplest way to use LGTM is as a Claude Code plugin. This gives you the /lgtm command, auto-starts the server, and connects MCP tools automatically.
In a Claude Code session, register the marketplace and install the plugin:
/plugin marketplace add tomplex/lgtm
/plugin install lgtm@tomplex-lgtm
Once installed, the server starts automatically when you open a Claude Code session (via a SessionStart hook). No manual MCP configuration needed - the plugin bundles its own .mcp.json. Just type /lgtm or ask Claude to review your changes.
For submitted reviews and direct questions to push back to Claude immediately (rather than requiring Claude to poll with read_feedback), start Claude Code with the development channels flag:
claude --dangerously-load-development-channels plugin:lgtm@tomplex-lgtm
Without this flag, everything still works - Claude just needs to call read_feedback to see your review comments.
/lgtm command - registers the current project and opens the review UI/lgtm analyze skill - dispatches agents to classify files by priority and generate a review strategy/lgtm walkthrough skill - generates a narrated tour of the diff as ordered "stops" you step through in the UI/lgtm prepare skill - chains analyze + walkthrough so a review session opens fully primedPATH — required for symbol lookup. Install with brew install ripgrep (macOS), apt install ripgrep (Debian/Ubuntu), or apk add ripgrep (Alpine). The server logs a prominent warning at startup if rg is missing, and symbol-lookup requests throw a clear error rather than returning empty results.Node dependencies are installed automatically on first session start. No manual build step needed.
You're working with Claude on a feature branch. When you're ready for review (or want Claude to review its own work), Claude registers the project with LGTM and opens a browser tab. The diff viewer shows your branch changes with syntax highlighting, word-level change detection, and inline commenting.
Claude can seed comments on specific lines before you even start reading. You reply, resolve, or dismiss those comments inline. When you're done, you submit your review - Claude reads the structured feedback and can act on it in the same session.
The review loop is iterative. Claude addresses feedback, you refresh the diff, post more comments. Back and forth until it's right.
Claude can also run an analysis pass that classifies every file in the diff by priority (critical / important / normal / low) and review phase (review / skim / rubber-stamp), then groups files thematically with a suggested review strategy. The sidebar switches between flat, grouped and phased views based on this analysis.

For larger or more architectural changes, Claude can generate a narrated walkthrough — an ordered sequence of "stops" through the diff, each with a title, narrative, and references to the specific lines and files it covers. You step through it in the UI like a guided tour. Stops are tracked as visited, the sidebar shows stop-coverage badges per file, and you can still leave comments and use peek-definition while inside walkthrough mode.
Press W from the diff to enter walkthrough mode. Enter and Shift+Enter step forward and back; g followed by a digit jumps directly to a stop; d exits back to the diff.

npx claudepluginhub tomplex/lgtm --plugin lgtmMulti-model consensus engine integrating OpenAI Codex CLI, Gemini CLI, and Claude CLI for collaborative code review and problem-solving.
Memory compression system for Claude Code - persist context across sessions
Standalone image generation plugin using Nano Banana MCP server. Generates and edits images, icons, diagrams, patterns, and visual assets via Gemini image models. No Gemini CLI dependency required.
Ultra-compressed communication mode. Cuts ~75% of tokens while keeping full technical accuracy by speaking like a caveman.
Comprehensive UI/UX design plugin for mobile (iOS, Android, React Native) and web applications with design systems, accessibility, and modern patterns
Write feature specs, plan roadmaps, and synthesize user research faster. Keep stakeholders updated and stay ahead of the competitive landscape.