Surgical Context MCP

surgical-context-mcp — Surgical AST Code Context & Knowledge Graph for AI Coding Agents

Stop feeding 3,000-line files into your LLM context window.

surgical-context-mcp is a zero-LLM-cost, embedded SQLite + Tree-sitter Code Knowledge Graph and MCP Server that gives AI coding agents (Claude Code, Cursor, Windsurf, Cline, Claude Desktop) the exact surgical context they need to edit, refactor, and navigate codebases safely — cutting token consumption by 70–95% without sacrificing code quality.

Real run on this repo: editing a function in a 1,264-line file costs ~3.7k tokens instead of ~13.3k (−72%) — and you still get the skeleton, callers and helpers.

⚡ Quick Start (30 seconds)

# 1. Install (or: pipx install / pip install surgical-context-mcp)

uv tool install surgical-context-mcp

2. Register with Claude Code

claude mcp add code-graph -- code-graph serve

3. Index your project + install the token-saving guard hook

code-graph install-hooks /path/to/your/project

That's it — your agent now calls get_edit_context instead of reading whole files. Using Cursor, Windsurf, Cline or Claude Desktop? See Step 1. Also listed on the official MCP Registry as io.github.erayerturk/surgical-context-mcp.

---

Table of Contents

1. Quick Start (30 seconds)
2. Why Use surgical-context-mcp?
3. How Much Does It Save & When Should You Use It?
4. Supported Languages & Ecosystem Resolvers
5. How It Works (Architecture)
6. Installation
7. How to Use (Step-by-Step Setup Guide)
8. Real Output Examples
9. MCP Tools Reference
10. CLI Reference
11. Recommended CLAUDE.md / .cursorrules Configuration

---

Why Use surgical-context-mcp?

The 3 Core Problems with AI Coding Agents Today

1. Full-File Reads Burn Hundreds of Thousands of Tokens: When you ask an AI agent (or when it spawns subagents) to fix a bug or add a feature, its default instinct is to run Grep and then Read entire 1,500–3,000 line files into the context window. Because every subsequent tool call re-sends the entire conversation history (cache reads), a single large file read early in a session compounds into 100k–300k wasted tokens.
2. Narrow Line-Range Reads Cause Hallucinations & Code Duplication: If you force an agent to read only lines 240–260 of a file to save tokens, the agent becomes blind to:

• Class attributes, self.x fields, @onready / @export variables, constants, and struct / interface definitions at the top of the file.
• Helper functions that already exist in the same file (causing the LLM to write duplicate code).
• How callers actually pass arguments to the function (±5 lines around call sites).
• The file's coding style (indentation, typing conventions, error handling idioms).

1. Pure Static AST Graphs Miss Dynamic Dispatch & Scene/Template Wiring: Standard call graphs only see static foo() calls. If a function is invoked via getattr(obj, "foo") in Python, call("foo") or a .tscn signal connection in Godot, SendMessage("foo") in Unity/C#, or referenced in a .vue/.html/.proto file, a naive graph reports *"0 callers — safe to delete"*, breaking production code.

How surgical-context-mcp Solves All Three

• `get_edit_context(symbol)` — One Call Before Any Edit: Instead of reading a 2,500-line file, the agent makes one MCP call (get_edit_context) and receives a strictly budgeted (≤ 4,000 tokens) surgical package containing:

1. Target Function Full Body (with exact line numbers for Edit)
2. File Skeleton (all imports, interface/struct/enum blocks, class fields, constants — without other function bodies)
3. Used Members & Direct Callees (definitions of class fields used by the target function + full bodies of short same-file helpers ≤ 25 lines)
4. Direct Callers (`d=1`, `±5` Line Windows) (exact call-site context without dumping 200-line caller functions)
5. Other Function Signatures in the File (so the agent reuses existing helpers instead of duplicating them)
6. Style Hints & Coverage Warnings (flags dynamic string calls, .tscn connections, templates, and protocol references)

• `check_rename_or_delete(symbol)` — One Call Before Any Refactor: Checks AST callers (exact/scoped), inheritance overrides across parent/child classes, .tscn/template connections, signal emit/connect sites, and workspace-wide dynamic string references.
• Cross-Language Contracts (`shares_contract`) — Server ↔ Client ↔ Config in One Call:
• Protocol Opcodes (`contract::opcode::0xCC`): Automatically links hex opcode constants across languages (e.g., Go OpSGMOnline = 0xCC and GDScript/TypeScript const S_GM_ONLINE = 0xCC), so querying either constant shows both server senders and client packet handlers together.
• Config & Schema Keys (`contract::config:: `): Automatically links Go struct json:"max_chars_per_ip" field tags, frontend/panel schema strings ('multi_guard.max_chars_per_ip'), and seed/*.json config entries ("max_chars_per_ip": 2) into a single shared contract node.
• Data, Seed & Translation (`i18n`) Graph (`check_translations`): Indexes seed/*.json, shared/maps/*.json, locales/*.json, and .po files alongside code. Links tr("...") and noticeT(..., "n.x") calls to locale definitions and audits missing keys, accidental raw non-key literals, %d/%s/{} placeholder count mismatches, and unused locale keys.
• Automated Test Discovery (`tests_for`) & Git Diff Blast Radius (`get_diff_impact`):
• tests_for(symbol) identifies every test exercising a symbol and returns the exact runnable CLI command (go test ./pkg -run '^TestFoo$', godot --headless --path client -s tests/..., pytest ... -k ...).
• get_diff_impact(base_ref="HEAD") parses git diff hunks, maps changed lines to modified AST symbols, and reports their callers, affected cross-language contracts, and runnable test commands in one call — built for code-review subagents (final-reviewer).
• Concept Search & Engine Lifecycle Noise Suppression:
• Indexes inline/preceding comments and docstrings (doc_comment) so searching by concept (e.g. "attack stat") finds KOAP (// attack power stat bonus).
• Recognizes engine-invoked virtual callbacks (_ready, _draw, _process, _physics_process, _input, Awake, Update) and suppresses false text-match noise warnings when AST upstream is empty.
• Direct Token Savings Telemetry (`get_token_savings`): Every MCP/CLI query records its returned token count vs. the raw token size of touched files in SQLite (usage_telemetry), answering *"How many tokens did code-graph save?"* with direct measurements instead of estimates.
• Zero LLM Indexing Cost & Sub-Second Speed: Built on native C Tree-sitter grammars and embedded SQLite (WAL mode). Indexes ~600 files in `< 0.6` seconds using 0 API tokens, and automatically refreshes modified files (mtime + SHA-1) on every query.
• Enforced via Claude Code `PreToolUse` Hook: Includes a built-in hook guard (code-graph install-hooks) that blocks unbounded `Read` calls on files `> 300` lines and returns the file's AST symbol table (Lstart-end: signature), ensuring both main agents and subagents stay on the surgical path.

---

How Much Does It Save & When Should You Use It?

1. Token & Cost Economics (Single-Turn vs. Multi-Turn Compounding Savings)

In agentic coding tools (Claude Code, Cursor, Cline), every tool call re-sends the entire conversation history up to that point. When an agent reads a 2,500-line file (~22,000 tokens) on Turn 2 of a 12-turn coding task, those 22,000 tokens are billed on Turn 2, Turn 3, Turn 4 ... all the way to Turn 12 (22,000 × 10 = 220,000 cumulative context tokens).

By replacing full-file reads with get_edit_context (~2,500–3,500 tokens), surgical-context-mcp prevents context bloat from compounding across turns.

Benchmark Across Common Developer Tasks

*(Measured on a 593-file multi-language repository: 2,607-line Godot client controller + Go backend + JS web dashboard)*

---

2. When Does surgical-context-mcp Provide the Biggest Gains?

High-Impact Scenarios (80%–97% Savings & Higher Code Accuracy)

• Large or Monolithic Source Files (`300` to `5,000+` lines): Game controllers (main.gd, PlayerController.cs), large backend services (`s