cortex-ast
A powerful Model Context Protocol (MCP) server and Omni-AST engine. It empowers AI agents to seamlessly parse complex codebases, execute secure cross-project operations, and dynamically fetch token-optimized rules.
Install / Use
claude mcp add cortex-works -- npx -y github:cortex-works/cortex-astIf the server publishes to npm under a different name, use that package instead — check the repo README.
MCP Server
Model Context Protocol server
Quality Score
Category
AI & Machine LearningSupported Platforms
Our assessment of cortex-ast
cortex-ast scores 78/100 on our quality scale, 796th of 947 AI & Machine Learning skills we index.
Its MCP Server is 6.7 KB long, well organised into 29 sections with 7 code examples: a thorough specification that gives an agent plenty to work with.
It has 10 GitHub stars, so there is little community track record yet; judge it on its content.
Maintenance, license and trust
- The repository was last updated about 9 months ago. That is recent enough to be usable, but agent tooling moves fast, so check the instructions against your agent's current version.
- It is released under the MIT license, a permissive license that allows use, modification and commercial use with attribution.
- Its trust signals score 91/100, with 1 caution from licensing, adoption, age or documentation. These come from repository metadata, not a code audit — read the skill file before letting an agent act on it.
Safety scan
No issues foundOur scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review of the same text found nothing harmful.
AI review by kimi-k2.7-code on 2026-11-14. Automated pattern scan on 2026-11-14. It catches known dangerous patterns, not every risk — read a skill before letting an agent act on it.
cortex-ast compared with similar skills
All 4 of these similar skills score higher than cortex-ast; compare them before choosing.
| Skill | Score | Stars | Updated | Format |
|---|---|---|---|---|
| cortex-ast (this skill)by cortex-works | 78 | 10 | 9mo ago | MCP Server |
| claude-memby thedotmack | 100 | 117.6k | today | CLAUDE.md |
| Agent-Reachby Panniantong | 100 | 112.3k | 8d ago | CLAUDE.md |
| Understand-Anythingby Egonex-AI | 100 | 92.8k | 1d ago | CLAUDE.md |
| rufloby ruvnet | 100 | 81.9k | today | CLAUDE.md |
Frequently asked questions
- How do I install cortex-ast?
- Run
claude mcp add cortex-works -- npx -y github:cortex-works/cortex-ast. The install tabs above show the steps for each supported agent. - Which AI agents does cortex-ast work with?
- It is written for Claude Code, Claude Desktop, Cursor and Zed, as a MCP Server file. Other agents that read the same format can often use it too.
- Is cortex-ast safe to use?
- Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review of the same text found nothing harmful. It is MIT-licensed and scores 91/100 on trust signals. Skills are instructions an agent will follow, so read the file before installing it and do not approve commands you do not understand.
- Is cortex-ast still maintained?
- The repository was last updated about 9 months ago. That is recent enough to be usable, but agent tooling moves fast, so check the instructions against your agent's current version.
Skill content
View source on GitHubCortexAST 🧠⚡
The AI-Native Code Intelligence Backend for LLM Agents Pure Rust · MCP Server · Semantic Code Navigation · AST Time Machine · Self-Evolving Wasm Parsers
What is CortexAST?
👁️ CortexAST is the "eyes" — read-only code intelligence. For write/execute capabilities, see the companion
cortex-actproject (the "hands").
CortexAST is a production-grade MCP (Model Context Protocol) server that gives AI coding agents (Claude, Gemini, GPT-4o, etc.) the ability to:
- Navigate codebases semantically — find symbols, blast-radius analysis, cross-file propagation checklists
- Evolve itself — download and hot-reload WebAssembly language parsers at runtime (Go, PHP, Ruby, Java, …)
- Time-travel your codebase — Chronos snapshot system for pre/post-refactor AST-level comparison
- Search local memory — hybrid semantic + keyword search over your codebase history
✋ To edit files, run commands, or patch configs, use cortex-act instead.
Feature Modules
1. 🔭 cortex_code_explorer
Codebase explorer. Use INSTEAD of ls/tree/find/cat. Two modes: map_overview (fast symbol map, near-zero tokens — run first on any repo) and deep_slice (token-budgeted XML with function bodies, vector-ranked by query). Use map_overview to orient; deep_slice to get code for editing.
2. 🎯 cortex_symbol_analyzer
AST symbol analysis. Use INSTEAD of grep/rg. Actions: read_source (extract exact source of a symbol from a file — do this before editing), find_usages (all call/type/field sites), find_implementations (structs implementing a trait), blast_radius (callers + callees — run before rename/delete), propagation_checklist (exhaustive update checklist for shared types).
3. ⏳ cortex_chronos
AST snapshot tool for safe refactors. Workflow: save_checkpoint (before edit) → edit → compare_checkpoint (verify). Use instead of git diff — AST-level, ignores formatting noise. Actions: save_checkpoint, list_checkpoints, compare_checkpoint, delete_checkpoint.
4. 🛠️ run_diagnostics
Run compiler diagnostics (cargo check / tsc / gcc). Call after any code edit to catch errors before proceeding. Returns file, line, code, message — structured for targeted fixes.
5. 🧠 cortex_memory_retriever
Search past agent decisions in global memory (semantic + keyword hybrid). Requires CortexSync. Call BEFORE any research or exploration — the answer may already be cached. Returns ranked entries: intent, decision, tags, files_touched.
6. 📋 cortex_get_rules
Fetch codebase AI rules for the current context. Requires CortexSync. Returns merged rules filtered by file_path (frontend/backend/db context). Call before starting any task in a new project.
7. ✨ cortex_remember
Save task outcome to permanent global memory. Requires CortexSync. Call at END of every task. intent+decision must be ≤200 chars each. For long artifacts write a file first and pass path via heavy_artifacts.
8. 🌍 cortex_list_network
List all AI-tracked codebases (Requires CortexSync network). Use to discover target_project IDs for cross-project operations.
9. 🌐 cortex_manage_ast_languages
Manage Wasm grammar parsers for non-core languages. Core (always active): rust, typescript, python. Call status to see active/available languages. Call add with languages[] to download and hot-reload parsers from GitHub tree-sitter releases. Available: go, php, cpp, c, c_sharp, java, ruby, dart.
Ecosystem Requirement: CortexSync 🧠
For full functionality, CortexSync (the "Brain") must be running in the background.
| Tool | Dependent on CortexSync? | Why? |
|---|---|---|
| cortex_remember | Yes | Persists task outcomes to the global journal. |
| cortex_memory_retriever| Yes | Performs semantic vector search over past decisions. |
| cortex_get_rules | Yes | Fetches centralized rules from the synchronized rule engine. |
| cortex_list_network | Yes | Reads the global network map of codebases. |
| cortex_code_explorer | No | Local AST analysis. |
| cortex_symbol_analyzer | No | Local AST analysis. |
If cortex-sync is offline, these tools will strictly return a graceful warning without interrupting the agent's workflow.
Quick Start
Prerequisites
- Rust 1.80+
- Ollama or LM Studio running locally (optional, for Auto-Healer)
Build & Run
git clone https://github.com/DevsHero/CortexAST
cd CortexAST
cargo build --release
# Run as MCP server (stdio)
./target/release/cortexast
MCP Config (~/.cursor/mcp.json or Claude Desktop)
{
"mcpServers": {
"cortexast": {
"command": "/path/to/cortexast",
"args": []
}
}
}
Usage Examples
Semantic Explorer — Bird's-eye view of a project
{
"name": "cortex_code_explorer",
"arguments": {
"action": "map_overview",
"target_dir": "."
}
}
Symbol Search — Find all usages across the repo
{
"name": "cortex_symbol_analyzer",
"arguments": {
"action": "find_usages",
"symbol_name": "AuthService",
"target_dir": "."
}
}
Time Travel — Compare AST after refactor
{
"name": "cortex_chronos",
"arguments": {
"action": "compare_checkpoint",
"symbol_name": "login",
"tag_a": "pre-refactor",
"tag_b": "__live__"
}
}
Self-Evolving Wasm Language Support
| Always Available | Downloadable on Demand | |---|---| | Rust, TypeScript/JS, Python | Go, PHP, Ruby, Java, C, C++, C#, Dart |
# Agent calls this automatically when it detects a new language:
cortex_manage_ast_languages { "action": "add", "languages": ["go", "dart"] }
Development
# Run all unit tests
cargo test
# Check (no link)
cargo check
---
## Architecture
CortexAST (binary) └── src/ ├── server.rs # MCP stdio server — all tool schemas + handlers ├── inspector.rs # LanguageConfig, LanguageDriver, Symbol, run_query ├── grammar_manager.rs # Wasm download + hot-reload (GitHub releases) ├── vector_store.rs # model2vec embeddings + cache invalidation ├── chronos.rs # AST snapshot time machine (Chronos) ├── memory.rs # global_memory.jsonl journal client └── project_map.rs # Network map for multi-repo roaming
## License
MIT — See [LICENSE](./LICENSE)
---
*Built with ❤️ in Rust · Semantic precision for the AI age*
Related Skills
claude-mem
117.6kPersistent Context Across Sessions for Every Agent – Captures everything your agent does during sessions, compresses it with AI, and injects relevant context back into future sessions. Works with Claude Code, OpenClaw, Codex, Gemini, Hermes, Copilot, OpenCode + More
Agent-Reach
112.3kGive your AI agent eyes to see the entire internet. Read & search Twitter, Reddit, YouTube, GitHub, Bilibili, XiaoHongShu — one CLI, zero API fees.
Understand-Anything
92.8kGraphs that teach > graphs that impress. Turn any code into an interactive knowledge graph you can explore, search, and ask questions about. Works with Claude Code, Codex, Cursor, Copilot, Gemini CLI, and more.
ruflo
81.9k🌊 The original agent harness. Deploy intelligent multi-player swarms, coordinate autonomous workflows, and build conversational AI systems. Features adaptive memory, self-learning intelligence, federation, vector RAG integration, and native Claude Code / Codex / Hermes and many more Integrated
Languages
Trust signals
From repository metadata: license, adoption, age and documentation. Not a code audit — see the Safety scan above for what the skill file itself contains.
