ucn
Universal Code Navigator - a lightweight MCP server that gives AI agents call-graph-level understanding of code. Instead of reading entire files, agents ask structural questions like: "who calls this function", "what breaks if I change it", "what's unused", and get precise, AST-verified answers. UCN parses JS/TS, Python, Go, Rust, Java, and HTML inline scripts with tree-sitter, then exposes 28 navigation commands as a CLI tool, MCP server, or agent skill.
UCN - Universal Code Navigator
See what code does before you touch it.
Find symbols, trace callers, check impact, pick the right tests, extract code and spot what's dead - from the terminal.
All commands, one engine, three surfaces:
Terminal AI Agents Agent Skills
│ │ │
CLI MCP Skill
└────────────────────┼────────────────────┘
│
┌──────┴──────┐
│ UCN Engine │
│ commands │
│ tree-sitter │
└─────────────┘
Supports JavaScript, TypeScript, Python, Go, Rust, Java, and HTML inline scripts.
If you work with AI, add UCN as a Skill or MCP and let the agent ask better code questions instead of reading whole files. All commands ship as a single tool.
UCN is deliberately lightweight:
- No background processes - parses on demand, answers, exits
- No language servers - tree-sitter does the parsing, no compilation needed
- MCP is optional - only needed if you connect UCN to an AI agent, the CLI and Skill work on their own
npm install -g ucn
ucn trace main --depth=3 # full execution flow
ucn about handleRequest # definition + callers + callees + tests
ucn impact handleRequest # every call site with arguments
ucn deadcode --exclude=test # unused code, AST-verified
"What happens when build() runs?"
$ ucn trace build --depth=2
build
├── detectProjectPattern (core/discovery.js:399) 1x
│ ├── checkDir (core/discovery.js:403) 2x
│ └── shouldIgnore (core/discovery.js:347) 1x
├── parseGitignore (core/discovery.js:130) 1x
├── expandGlob (core/discovery.js:190) 1x
│ ├── parseGlobPattern (core/discovery.js:226) 1x
│ ├── walkDir (core/discovery.js:283) 1x
│ └── compareNames (core/discovery.js:169) 1x
├── indexFile (core/project.js:273) 1x
│ ├── addSymbol (core/project.js:343) 4x
│ ├── detectLanguage (languages/index.js:157) 1x
│ ├── parse (core/parser.js:69) 1x
│ └── extractImports (core/imports.js:19) 1x
├── buildImportGraph (core/project.js:549) 1x
└── buildInheritanceGraph (core/project.js:627) 1x
One command. No files opened. Every function located by file and line.
Understand code you didn't write
ucn about gives you everything about a function in one shot - who calls it, what it calls, which tests cover it, and the source code.
$ ucn about expandGlob
expandGlob (function)
════════════════════════════════════════════════════════════
core/discovery.js:190-221
expandGlob (pattern, options = {})
CALLERS (3):
cli/index.js:859 [runGlobCommand]
const files = expandGlob(pattern);
core/cache.js:274 [isCacheStale]
const currentFiles = expandGlob(pattern, globOpts);
core/project.js:195 [build]
const files = expandGlob(pattern, globOpts);
CALLEES (3):
parseGlobPattern [utility] - core/discovery.js:226
walkDir [utility] - core/discovery.js:283
compareNames [utility] - core/discovery.js:169
TESTS: 6 matches in 2 file(s)
Need to trace execution upward instead? ucn reverse-trace fn walks the caller chain back to entry points.
Change code without breaking things
Before touching a function, check if all existing call sites match its signature:
$ ucn verify expandGlob
expandGlob (pattern, options = {})
Expected arguments: 1-2
STATUS: ✓ All calls valid (7 calls, 0 mismatches)
Then preview the refactoring. UCN shows exactly what needs to change and where:
$ ucn plan expandGlob --rename-to=expandGlobPattern
SIGNATURE CHANGE:
Before: expandGlob (pattern, options = {})
After: expandGlobPattern (pattern, options = {})
CHANGES NEEDED: 7 across 4 files
cli/index.js :859
const files = expandGlob(pattern);
→ const files = expandGlobPattern(pattern);
core/cache.js :274
const currentFiles = expandGlob(pattern, globOpts);
→ const currentFiles = expandGlobPattern(pattern, globOpts);
core/project.js :195
const files = expandGlob(pattern, globOpts);
→ const files = expandGlobPattern(pattern, globOpts);
Run ucn diff-impact --staged before committing to see what you changed and who calls it.
Find what to clean up
Which tests should you run after a change? affected-tests walks the blast radius and finds every test that touches the affected functions:
$ ucn affected-tests expandGlob
1 function changed → 18 functions affected (depth 3)
Summary: 18 affected → 12 test files, 11/18 functions covered (61%)
Uncovered (7): runGlobCommand, runProjectCommand, ...
⚠ These affected functions have no test references
Find unused code
$ ucn deadcode --exclude=test
Dead code: 1 unused symbol(s)
core/discovery.js
[ 162- 166] legacyResolve (function)
Extract without reading the whole file
$ ucn fn compareNames
core/discovery.js:169
[ 169- 177] compareNames(a, b)
────────────────────────────────────────────────────────────
function compareNames(a, b) {
const aLower = a.toLowerCase();
const bLower = b.toLowerCase();
if (aLower < bLower) return -1;
if (aLower > bLower) return 1;
if (a < b) return -1;
if (a > b) return 1;
return 0;
}
Testing and reliability
- Fast - uses incremental cache for optimal performance
- Discipline - every bug fix gets a regression test, test code is ~3x the source
- Coverage - every command, every supported language, every surface (CLI, MCP, interactive)
- Systematic - a harness exercises all command and flag combinations against real multi-language fixtures
- Test types - unit, integration, per-language regression, formatter, cache, MCP edge cases, architecture parity guards
AI Setup
MCP
# Claude Code
claude mcp add ucn -- npx -y ucn --mcp
# OpenAI Codex CLI
codex mcp add ucn -- npx -y ucn --mcp
# VS Code Copilot
code --add-mcp '{"name":"ucn","command":"npx","args":["-y","ucn","--mcp"]}'
Or add to the MCP config file manually
{
"mcpServers": {
"ucn": {
"command": "npx",
"args": ["-y", "ucn", "--mcp"]
}
}
}
VS Code uses .vscode/mcp.json:
{
"servers": {
"ucn": {
"type": "stdio",
"command": "npx",
"args": ["-y", "ucn", "--mcp"]
}
}
}
Agent Skill (no server needed)
# Claude Code
mkdir -p ~/.claude/skills
cp -r "$(npm root -g)/ucn/.claude/skills/ucn" ~/.claude/skills/
# OpenAI Codex CLI
mkdir -p ~/.agents/skills
cp -r "$(npm root -g)/ucn/.claude/skills/ucn" ~/.agents/skills/
Full help
UCN - Universal Code Navigator
Supported: JavaScript, TypeScript, Python, Go, Rust, Java, HTML
Usage:
ucn [command] [args] Project mode (current directory)
ucn <file> [command] [args] Single file mode
ucn <dir> [command] [args] Project mode (specific directory)
ucn "pattern" [command] [args] Glob pattern mode
(Default output is text; add --json for machine-readable JSON)
═══════════════════════════════════════════════════════════════════════════════
UNDERSTAND CODE
═══════════════════════════════════════════════════════════════════════════════
about <name> Full picture (definition, callers, callees, tests, code)
context <name> Who calls this + what it calls (numbered for expand)
smart <name> Function + all dependencies inline
impact <name> What breaks if changed (call sites grouped by file)
blast <name> Transitive blast radius (callers of callers, --depth=N)
trace <name> Call tree visualization (--depth=N expands all children)
reverse-trace <name> Upward call chain to entry points (--depth=N, default 5)
related <name> Find similar functions (same file, shared deps)
example <name> Best usage example with context
═══════════════════════════════════════════════════════════════════════════════
FIND CODE
═══════════════════════════════════════════════════════════════════════════════
find <name> Find symbol definitions (supports glob: find "handle*")
usages <name> All usages grouped: definitions, calls, imports, references
toc Table of contents (compact; --detailed lists all symbols)
search <term> Text search (regex default, --context=N, --exclude=, --in=)
Structural: --type=function|class|call --param= --returns= --decorator= --exported --unused
tests <name> Find test files for a function
affected-tests <n> Tests affected by a change (blast + test detection, --depth=N)
═══════════════════════════════════════════════════════════════════════════════
EXTRACT CODE
═══════════════════════════════════════════════════════════════════════════════
fn <name>[,n2,...] Extract function(s) (comma-separated for bulk, --file)
class <name> Extract class
lines <range> Extract line range (e.g., lines 50-100)
expand <N> Show code for item N from context output
═══════════════════════════════════════════════════════════════════════════════
FILE DEPENDENCIES
═══════════════════════════════════════════════════════════════════════════════
imports <file> What does file import
exporters <file> Who imports this file
file-exports <file> What does file export
graph <file> Full dependency tree (--depth=N, --direction=imports|importers|both)
circular-deps Detect circular import chains (--file=, --exclude=)
═══════════════════════════════════════════════════════════════════════════════
REFACTORING HELPERS
═══════════════════════════════════════════════════════════════════════════════
plan <name> Preview refactoring (--add-param, --remove-param, --rename-to)
verify <name> Check all call sites match signature
diff-impact What changed in git diff and who calls it (--base, --staged)
deadcode Find unused functions/classes
entrypoints Detect framework entry points (routes, DI, tasks)
═══════════════════════════════════════════════════════════════════════════════
OTHER
═══════════════════════════════════════════════════════════════════════════════
api Show exported/public symbols
typedef <name> Find type definitions
stats Project statistics (--functions for per-function line counts)
stacktrace <text> Parse stack trace, show code at each frame (alias: stack)
Common Flags:
--file <pattern> Filter by file path (e.g., --file=routes)
--exclude=a,b Exclude patterns (e.g., --exclude=test,mock)
--in=<path> Only in path (e.g., --in=src/core)
--depth=N Trace/graph depth (default: 3, also expands all children)
--direction=X Graph direction: imports, importers, or both (default: both)
--all Expand truncated sections (about, trace, graph, related)
--top=N Limit results (find, deadcode)
--limit=N Limit result count (find, usages, search, deadcode, api, toc)
--max-files=N Max files to index (large projects)
--context=N Lines of context around matches
--json Machine-readable output
--code-only Filter out comments and strings
--with-types Include type definitions
--include-tests Include test files
--class-name=X Scope to specific class (e.g., --class-name=Repository)
--include-methods Include method calls (obj.fn) in caller/callee analysis
--include-uncertain Include ambiguous/uncertain matches
--no-confidence Hide confidence scores (shown by default)
--min-confidence=N Filter edges below confidence threshold (0.0-1.0)
--include-exported Include exported symbols in deadcode
--no-regex Force plain text search (regex is default)
--functions Show per-function line counts (stats command)
--include-decorated Include decorated/annotated symbols in deadcode
--framework=X Filter entrypoints by framework (e.g., --framework=express,spring)
--exact Exact name match only (find)
--calls-only Only show call/test-case matches (tests)
--case-sensitive Case-sensitive text search (search)
--detailed List all symbols in toc (compact by default)
--top-level Show only top-level functions in toc
--max-lines=N Max source lines for class (large classes show summary)
--no-cache Disable caching
--clear-cache Clear cache before running
--base=<ref> Git ref for diff-impact (default: HEAD)
--staged Analyze staged changes (diff-impact)
--no-follow-symlinks Don't follow symbolic links
-i, --interactive Keep index in memory for multiple queries
Limitations
- Single-project scope - follows imports within the project, not into
node_modulesorsite-packages - No runtime execution - static analysis only
- Dynamic dispatch and reflection are only partially visible or invisible
- JS, TS, and Python method calls can be uncertain when receiver type is unknown
- Large repos take a few seconds on the first query, then use cache
If you need compiler diagnostics, taint analysis, or runtime semantics, those are different tools for different jobs. UCN trades that depth for speed, portability, and zero setup.
MIT
相关服务器
Scout Monitoring MCP
赞助Put performance and error data directly in the hands of your AI assistant.
Alpha Vantage MCP Server
赞助Access financial market data: realtime & historical stock, ETF, options, forex, crypto, commodities, fundamentals, technical indicators, & more
cratesio-mcp
MCP server for querying crates.io - the Rust package registry
Postman MCP Server
Run Postman collections using Newman, with support for environment and global variables.
Remote MCP Server (Authless)
A template for deploying a remote, auth-less MCP server on Cloudflare Workers.
gopls-mcp
The essential MCP server for Go language: Exposing compiler-grade semantics to AI Agents and LLM for deterministic code analysis and minimal token usage.
Verilator MCP Server
An MCP server for Verilator providing RTL simulation, automatic testbench generation, and natural language query capabilities.
JetBrains
Work on your code with JetBrains IDEs
Copado MCP
A remote MCP server deployable on Cloudflare Workers without authentication.
MCP Bench Router
Claude Code sucks at design. Let it delegate it's tasks to better models. Claude will use the MCP to get leaderboard of best design models and query specific code changes using OpenRouter.
BlenderMCP
Connects Blender to Claude AI via the Model Context Protocol (MCP), enabling direct AI interaction for prompt-assisted 3D modeling, scene creation, and manipulation.
ImageJ / Fiji
An MCP server for ImageJ/Fiji, implemented as a Python wrapper for Fiji functions.