
ariadne
io.github.MikhailHalPublished by GitHub account MikhailHalv0.4.0Updated Oct 24, 2026
Returns only the unit tests affected by your current Kotlin/Android code changes.
Overview
An MCP server that analyzes Kotlin/Android code changes and returns only the unit tests affected by them.
- What you can do
- Ariadne exposes a get_affected_tests tool that takes a Kotlin project path and returns a sorted list of affected test fully qualified names. It runs git diff internally against a base branch or HEAD, then uses the sazanami Kotlin Analysis API to build a call graph and find affected tests. It reports explicitly when changes fall outside analyzed sources, such as build scripts or resources, and returns an error on its 120-second analysis timeout.
- When to use it
- Useful during an agent's edit-verify loop on Kotlin or Android projects, where running only the affected tests gives faster feedback than the full suite. It is not a replacement for CI: the README advises keeping a full or conservative test run before merging.
- Requirements
- Runs as a local process over stdio. Needs JDK 21 or later and Git for diff detection. Installable via Homebrew, a Docker image, or a release JAR; the Docker route requires mounting the project and passing /workspace as project_path. No accounts, API keys, or environment variables are declared.
Installation
In SourceWeft
- Open ariadne in the dashboard and add it to a workspace.
- Enable the server for the chats that should use its tools.
Desktop only via STDIO. STDIO servers start a local process, so they need the SourceWeft desktop host.
Other MCP clients
Follow the launch instructions in the repository.
README
ariadne
MCP Server for Affected Test Selection
[Release] [MCP Registry] [Homebrew] [Container image] [License] [Kotlin]
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]
⚠️ MUST READ: Speed over Completeness
ariadne is built for the agent inner loop — edit, verify, commit — 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 — Works with Claude Code, Claude Desktop, and other MCP-compatible clients
- Automatic Git Diff — No need to pass diff manually; ariadne runs
git diffinternally - Powered by sazanami — Uses Kotlin Analysis API for accurate static analysis
Installation
Homebrew (recommended)
Then register it with your MCP client — for Claude Code:
Or add to your MCP client configuration manually (e.g., Claude Desktop):
Docker
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
Usage
Once configured, AI agents can use the get_affected_tests tool:
Tool: get_affected_tests
Parameters:
project_path(required) — Path to the Kotlin projectscope(optional,deep|shallow, defaultdeep) — how far back to look:deep— 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.shallow— 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) — 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:
Response:
How It Works
- Agent calls tool — Passes project path to ariadne
- Run git diff — ariadne executes
git diff --unified=0against base branch - Analyze with sazanami — Build call graph and find affected tests
- Return results — List of test FQNs returned to agent
Real-World Validation: Now in Android
Measured against Now in Android (Google's reference Android app — 34 modules, ~268 Kotlin files):
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:
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 — plan your CI safety net around them:
Full audit notes: sazanami#29, sazanami#38.
Requirements
- JDK 21 or later
- Git — 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 — 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
See LICENSE for the full text.
Source: README.md at commit 057144e
Tools
0Version history
1- v0.4.0LatestSep 16, 2026
Score
50% of the applicable checks met
How complete this listing is, computed from the registry entry, the endpoint and public sources. Not a rating of the server and not a statement that it is safe or working today: reachability is a dated check, and no installs or reviews go into the score.
- Reached in the last 30 days25 pts
Keep the endpoint answering the MCP handshake; SourceWeft probes it every 30 days.
- Auth declared10 pts
Declare the credential headers in the registry entry, or keep the endpoint answering the probe so its auth can be classified.
- Tools listed15 pts
Answer tools/list without a credential, or ship a package that starts in a sandbox without one.
- DescriptionRequired15 pts
- Logo10 pts
- Documentation15 pts
- License10 pts