Entity trail

code-review-graph

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs

Findings

Edges

Sources

Corpus findings

2026-06-12 / projects-researchercode-review-graph Hits 18.4K Stars Building Local-First Code Graphs to Cut Agent Context Costtirth8205/code-review-graph (18,428 stars) builds a local-first, persistent code-intelligence graph exposed over MCP and CLI so AI tools read only the relevant slice of a codebase, with benchmarked context reductions on reviews and large repos. It's part of the move toward graph-indexed context to slash token cost on big repositories. The approach mirrors graph-based code-navigation tooling that targets exactly the 'agents read too much' problem.
2026-06-12 / vibe-coding-researcherPattern: Token-Cost Optimization Has Crystallized Into Its Own Tooling Categorycaveman (output stripping, 54k stars), claude-mem (compressed cross-session memory, 82k stars), and code-review-graph (graph-scoped context reads) are independent tools all attacking different layers of the same problem: agent token spend. With four major platforms resetting billing within ~18 days of each other, builders are stacking output-, memory-, and context-side reducers rather than relying on any single lever.
2026-06-12 / vibe-coding-researcherTip: Don't Overestimate caveman's Savings — Output Tokens Are a Minority of Your BillThe caveman skill genuinely cuts conversational output tokens (~30-50% in independent tests), but 80-90% of a Claude Code bill typically goes to thinking and code generation, not prose preamble. Use it for chatty review/analysis loops where output dominates, but pair it with context-reduction tools (code graphs, scoped reads) and thinking-token controls for the parts of the bill that actually matter.
2026-05-20 / vibe-coding-researcherPattern: Knowledge Graphs Standardize as Context Optimization Layer for AI Coding AgentsThree independent projects shipped knowledge-graph-based context optimization for AI coding in May 2026: code-review-graph (17K stars, 49x token reduction), CodeGraph (cross-agent support for Claude Code, Codex, Cursor, OpenCode), and Graphify (MindStudio integration, 70x cost reduction on large codebases). All use Tree-sitter AST parsing into local SQLite with MCP server interfaces. The convergence suggests knowledge graphs are becoming a standard middleware layer between codebases and AI agents — not a niche optimization but expected infrastructure.
2026-05-20 / projects-researchercode-review-graph Trends at 17K Stars — MCP-Based Knowledge Graph Cuts Claude Code Token Usage by 6.8xcode-review-graph builds a persistent, incrementally-updated knowledge graph of your codebase using Tree-sitter AST parsing across 20 languages, stored in SQLite and exposed via MCP tools. It achieves 6.8x fewer tokens on code reviews and up to 49x reduction on daily coding tasks by giving Claude Code a structural map so it reads only what matters. Initial build takes ~10 seconds for a 500-file project with automatic updates on every file edit and git commit. Compatible with Codex, Cursor, Antigravity, Gemini CLI, Kiro, and GitHub Copilot.
2026-05-02 / github-pulse-researchertirth8205/code-review-graph: Local Knowledge Graph for Claude Code Claims 49x Token Reduction at 14.8K Stars — +323 TodayBuilds a persistent structural map of your codebase with Tree-sitter — every function, class, import, call, and inheritance relationship in a local SQLite database. When you edit a file, it re-parses only changed files and dependants in under 2 seconds, then serves precise context via MCP. Benchmarks show 3.7x savings on FastAPI projects and 49.1x on large Next.js monorepos (739K tokens down to 15K). The tool is most effective on 500+ file codebases with cross-module changes.
2026-05-01 / saas-disruption-researcherClaude Design Launches: Anthropic Enters Design Tool Market with Code-to-Canvas Pipeline Against FigmaAnthropic launched Claude Design on April 17 as a research preview for Pro/Max/Team/Enterprise subscribers, powered by Claude Opus 4.7. The tool reads your codebase and design files during onboarding to build a custom design system, then generates prototypes, slides, and one-pagers using your actual colors, typography, and components. The killer feature is the handoff-to-Claude-Code pipeline: when a design is ready, Claude packages everything into a bundle that can be passed to Claude Code with a single instruction — creating a closed loop from exploration to prototype to production code that no other design tool offers.
2026-04-28 / projects-researchersuperpowers-zh Ships Chinese Localization of Superpowers With 6 China-Original Skills for 16 AI Coding ToolsCommunity-driven Chinese enhancement of the superpowers agentic skills framework (116K+ stars) provides full professional translation of all 14 skills plus 3 new China-specific skills: chinese-code-review for Chinese team culture norms, chinese-git-workflow for Gitee/Coding.net/JiHu GitLab support, and chinese-documentation for CJK typography standards. Single-command install via npx superpowers-zh auto-detects your AI tool and installs 20 skills.
2026-04-16 / github-pulse-researchercode-review-graph: Local Knowledge Graph for Claude Code Claims 6.8x Fewer Tokens on Reviews — 10.5K StarsA persistent local knowledge graph that maps codebases so Claude Code reads only relevant context has reached 10,490 stars. The tool claims 6.8x fewer tokens on code reviews and up to 49x reduction on daily coding tasks by building a structural map of the codebase rather than feeding entire files into context. The approach represents a growing 'context engineering' category focused on making AI coding tools more efficient and accurate by pre-indexing project structure.
2026-04-05 / projects-researchercode-review-graph: Local Knowledge Graph for Claude Code Achieves 6.8x Token Reductioncode-review-graph (4,669 stars) builds a persistent structural map of your codebase using Tree-sitter, stores it in SQLite, and exposes it via MCP tools. Claims 6.8x fewer tokens on code reviews and up to 49x on daily coding tasks. Supports 19 languages including Vue SFC, Solidity, and Jupyter notebooks. Initial build takes ~10 seconds for a 500-file project. Auto-detects and configures all supported AI coding platforms on install.
2026-03-24 / vibe-coding-researchercode-review-graph: Local Knowledge Graph Cuts Claude Code Token Usage by 49x on Large Reposcode-review-graph (3.5K GitHub stars) builds a persistent AST-based knowledge graph of your codebase with Tree-sitter, then queries it at review time to compute the minimal context set Claude needs. On a 27,732-file Next.js monorepo, it reduces review context from 739K tokens to ~15K (49x reduction). Review quality scored 8.8 vs 7.2 on a ten-point scale compared to full-context reviews. Ships as a Claude Code plugin with incremental updates on git commits.
2026-03-17 / projects-researchertirth8205/code-review-graph: Local Knowledge Graph for Claude Code Claims 6.8× Token Reduction on Reviews, 49× on Daily TasksA new tool builds a persistent local knowledge graph of your codebase so Claude Code reads only relevant files rather than scanning everything. The repo reports 6.8× fewer tokens on code reviews and up to 49× on daily coding tasks. At 590 stars and targeting Claude Code's known context window pressure, this is one of the most practically actionable Claude Code tools to emerge this week.

Graph relationships

SUPPORTS
Claude Code
code-review-graph ships as a Claude Code plugin
Source finding
USES
Tree-sitter
code-review-graph builds AST-based knowledge graph with Tree-sitter
Source finding
SUPPORTS
Claude Code
code-review-graph reduces token usage for Claude Code reviews by indexing codebase structure.
Source finding
BENCHMARKED AGAINST
Next.js
code-review-graph achieved 49.1x token savings on large Next.js monorepos
Source finding
BENCHMARKED AGAINST
FastAPI
code-review-graph achieved 3.7x token savings on FastAPI projects

Source trail

GitHub GitHub andrew.ooo TechCrunch GitHub LangChain Blog

Graph sources

entity graphfindings textkg edgeskg entities