MikhailHal/ariadne
ā š š šŖ š§ - Affected-test selection for Kotlin/Android projects: returns only the unit tests impacted by the current working-tree changes, so agents skip full test runs while iterating. Function-level static analysis via Kotlin Analysis API ā call graph over functions, properties (initializers, delegates, accessors), constructors and callable references ā with automatic multi-module and Android variant source-set discovery. Verified on Now in Android (34 modules, 4s per query). brew install mikhailhal/tap/ariadne
Quick Install
{
"mcpServers": {
"mikhailhal-ariadne": {
"command": "npx",
"args": [
"-y",
"mikhailhal-ariadne"
]
}
}
}Using an AI coding agent (Claude Code, Cursor, etc.)? Copy a ready-made prompt that tells it to fetch the setup instructions and install this server for you.
Documentation Overview
ariadne
MCP Server for Affected Test Selection
ariadne is an MCP (Model Context Protocol) server that provides AI agents with the ability to identify affected tests. Powered by sazanami, it analyzes code changes and returns only the tests that need to be run.
[!IMPORTANT]
ā ļø MUST READ: Speed over Completeness
ariadne is built for the agent inner loop ā edit, verify, commit ā where fast feedback matters more than exhaustive selection. Static analysis cannot trace every execution path: reflection, DI frameworks, and data-flow indirection (e.g., Flux/MVI dispatch) can hide dependencies from any affected-test-selection tool, not just ariadne.
Always keep a final line of defense in CI. Run the full test suite (or a conservative selection) before merging. ariadne narrows what an agent runs while iterating; it is not a replacement for CI.
When ariadne detects changes it cannot analyze (build scripts, resources, unscanned source sets), it says so explicitly in the tool response instead of silently reporting "no affected tests".
Features
- MCP Integration ā Works with Claude Code, Claude Desktop, and other MCP-compatible clients
- Automatic Git Diff ā No need to pass diff manually; ariadne runs
git diffinternally - Powered by sazanami ā Uses Kotlin Analysis API for accurate static analysis
Installation
Homebrew (recommended)
brew install mikhailhal/tap/ariadne
Then register it with your MCP client ā for Claude Code:
claude mcp add ariadne -- ariadne
Or add to your MCP client configuration manually (e.g., Claude Desktop):
{
"mcpServers": {
"ariadne": {
"command": "ariadne"
}
}
}
Docker
docker run -i --rm -v /path/to/project:/workspace ghcr.io/mikhailhal/ariadne
Mount the project you want analyzed and pass /workspace as project_path.
The image is also listed in the official MCP Registry as io.github.MikhailHal/ariadne.
Manual (release JAR)
Download ariadne-<version>-all.jar from Releases (requires JDK 21+) and configure your client with "command": "java", "args": ["-jar", "/path/to/ariadne-<version>-all.jar"].
Build from Source
git clone --recursive https://github.com/MikhailHal/ariadne.git
cd ariadne
./gradlew shadowJar # fat JAR: build/libs/ariadne-<version>-all.jar
Usage
Once configured, AI agents can use the get_affected_tests tool:
Tool: get_affected_tests
Parameters:
project_path(required) ā Path to the Kotlin projectscope(optional,deep|shallow, defaultdeep) ā how far back to look:deepā all changes sincebase_branch(committed and uncommitted). Safest; the whole branch is covered so nothing you already committed slips through unverified. May select more tests.shallowā uncommitted changes only (diff againstHEAD). Fastest, for the tight edit loop. Verifying already-committed work is left to the caller.base_branchis ignored.
base_branch(optional,deepscope only) ā Branch to compare against. When omitted, ariadne uses the repository's default branch (origin/HEAD). If that is not set (e.g. a repo with no remote), it returns an error asking you to passbase_branchexplicitly rather than guessing.
Returns:
- List of affected test FQNs (fully qualified names), sorted
- If the diff contains changes outside the analyzed Kotlin sources (build scripts, resources, unscanned source sets), a note is appended recommending a full test run for those changes
- Analysis is bounded by a 120s timeout; on timeout an explicit error is returned
Example
Agent request:
{
"name": "get_affected_tests",
"arguments": {
"project_path": "/path/to/your/kotlin/project"
}
}
Response:
com.example.UserServiceTest.testCreateUser
com.example.UserRepositoryTest.testSave
How It Works
āāāāāāāāāāāāāāā āāāāāāāāāāāāāāā āāāāāāāāāāāāāāā
ā MCP Client ā āāā¶ ā ariadne ā āāā¶ ā sazanami ā
ā (Agent) ā ā (MCP Server)ā ā (Analysis) ā
āāāāāāāāāāāāāāā āāāāāāāāāāāāāāā āāāāāāāāāāāāāāā
ā
ā¼
āāāāāāāāāāāāāāā
ā git diff ā
āāāāāāāāāāāāāāā
- Agent calls tool ā Passes project path to ariadne
- Run git diff ā ariadne executes
git diff --unified=0against base branch - Analyze with sazanami ā Build call graph and find affected tests
- Return results ā List of test FQNs returned to agent
Real-World Validation: Now in Android
Measured against Now in Android (Google's reference Android app ā 34 modules, ~268 Kotlin files):
| Metric | Result |
|---|---|
| Recall audit ā 19 target functions across all layers | 18/18 valid targets detected (the 19th had no exercising unit test; correctly not selected) |
| End-to-end response time | ~4s (module discovery + call-graph build + BFS) |
| Module discovery | 34 modules via settings.gradle.kts, incl. nested modules and type-safe accessor dependencies |
| Source sets | main, debug, prod, benchmark, testDemo, ⦠discovered per module (androidTest* excluded by design) |
Verified patterns include repositories behind project interfaces, a library-interface
override (androidx.datastore.Serializer), operator fun invoke use cases,
@Composable functions, extension mappers, ViewModel property-initializer chains,
and callable references. Two representative results:
- Changing
core:common'sasResult()selects 14 tests across three modules, including ViewModel tests reachable only throughval uiState = ...stateIn(...) - Changing the mapper
PopulatedNewsResource.asExternalModel()selects 14 tests, including 11 repository tests reachable only through.map(Type::mapper)chains
Test-class selection rate
Every unit-test class in Now in Android was measured by changing a function in the class it tests and checking whether that test class was selected:
| Test style | Selected |
|---|---|
| Plain unit tests (construct the object, call it) | 13 / 13 valid targets |
| Robolectric / Compose screenshot tests | 12 / 12 |
Framework-dispatched callbacks (lint Detector) | 0 / 2 ā see below |
Robolectric turned out not to be a barrier: those tests call the composable
themselves (setContent { NiaTheme { ... } }), so the call exists in the source.
What decides coverage is not the test runner but whether the test's own code
contains the call.
What ariadne cannot see
The rule of thumb: if the framework calls your code instead of your test calling it, ariadne cannot connect them. These are limits of static analysis, not bugs ā plan your CI safety net around them:
| Pattern | Status |
|---|---|
Framework-invoked callbacks ā Fragment/Activity lifecycle (launchFragmentInContainer), lint Detector methods, Application.onCreate | Not traced: no call written in the test |
| Reflection / DI-container wiring | Not traced |
| UDF dispatch (Flux/MVI) | dispatch ā collect is never an edge, but wiring in init (or a start() the test calls) is covered conservatively via constructor chains. Subscriptions started by DI/lifecycle are not covered (sazanami#38) |
stateIn / shareIn chains (map, onEach, flatMapLatest, combine) | Covered ā verified with exact selection |
Instrumented tests (androidTest*) | Out of scope by design |
| Build scripts, resources, unscanned source sets | Not analyzed ā reported explicitly in the tool response |
| Same-name top-level extensions in one package | Over-selected (receiver types are not part of top-level FQNs) ā safe direction |
KMP source sets (commonMain, expect/actual) | Enumerated, but resolution quality unverified (#1) |
Full audit notes: sazanami#29, sazanami#38.
Requirements
- JDK 21 or later
- Git ā For diff detection
Limitations
- Module discovery is convention-based: it parses
settings.gradle(.kts)includes, enumeratessrc/<sourceSet>/{kotlin,java}layouts, and readsproject(":x")/ type-safe accessor dependencies from build files. Dynamic includes,projectDirremapping, customsrcDirs, and dependencies injected by convention plugins are not detected ā see #1 - Full graph rebuild on each request (no caching yet); analysis is capped at 120s
- See What ariadne cannot see for analysis-level gaps
License
Copyright 2025 ariadne contributors
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
See LICENSE for the full text.