Engine Atlas MCP
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@Engine Atlas MCPStudy the class gin::Widget in Engine Atlas."
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Engine Atlas MCP
Research the Asphalt 8 game engine with your AI assistant. Engine Atlas MCP gives Claude Code, Codex, Antigravity or any other MCP client the tools of the Engine Atlas reference: the class catalog, decompilation through Ghidra, notes, card writing and quality checks.
Everything is saved to local drafts. Nothing reaches the public reference until you send it and it passes review.
What you need
Python 3.11 or newer.
A database snapshot from the Engine Atlas maintainer. It is not distributed with this project, and the server cannot start without it.
For decompilation: Ghidra with Ghidra MCP, and
libasphalt8.sofrom the Engine Atlas site, opened in Ghidra.
Related MCP server: saripati
Install
pip install git+https://github.com/mindkeyboard/engine-atlas-mcp
engine-atlas-mcp init --snapshot atlas-snapshot.sqlite
engine-atlas-mcp doctor --elf /path/to/libasphalt8.soData goes to ~/.engine-atlas. Use --data, or the ENGINE_ATLAS_DATA variable, to
keep it elsewhere.
Then connect it to your client: Claude Code, Codex and Antigravity.
Use
Ask the model to study a class:
Study the class gin::Widget in Engine Atlas.
The model reads the full research rules from the server itself, so there is nothing else to teach it. Strong models are strongly recommended, such as Sonnet 5.5, GPT-6 Sol or newer. AI research is valued above manual work because its whole process stays visible, so keep the notes.
Before you send anything, have a second session check the result as a critic:
You are the critic. Check the cards for gin::Widget by the critic role in the guide.
When the critic has no findings left:
Export my drafts with atlas_export_drafts.
Open My edits on the site, choose Load a file, review the edits and send them.
What is different from the maintainer's copy
Tools work the same, on your local copy of the database. Four publication actions are
absent: atlas_promote_class_publication, atlas_promote_class_contract,
atlas_waive_class_method and atlas_refresh_reference. Two tools are added:
atlas_list_drafts and atlas_export_drafts.
The export carries articles, method descriptions, properties, examples, evidence and notes. Method order, class contracts and executable bindings stay local for now.
How it fits together
src/engine_atlas holds the parts of the research core that the server uses.
src/engine_atlas_mcp adds local drafts and the export. The original game file is
never changed, and no game files are part of this repository.
License
Apache-2.0. This is an independent project, not affiliated with Gameloft. See NOTICE.
This server cannot be deployed
Maintenance
Related MCP Connectors
- HiveHallOAuthai.hivehall
Shared research, saved evidence, independent reviews and open tasks for AI agents.
Agent-native notes, tasks, dev-docs, vaults, sync & handoffs. MCP + OpenAPI dual surface.
Gives your AI assistant persistent memory and intelligence about your work patterns.
Markdown notes and whiteboards your AI agent can read and write.
Related MCP Servers
- AlicenseNot gradedqualityDmaintenanceTransforms AI assistants into research-grade cognitive workspaces with systematic reasoning, evidence-based analysis, persistent memory management, and intelligent knowledge discovery.MIT
- AlicenseNot gradedqualityAmaintenanceProvides AI assistants a local, hybrid semantic+keyword memory vault to save research findings, notes, and project context, and recall them across sessions with session boot/save tools.4 npm2MIT
- AlicenseNot gradedqualityCmaintenanceProvides local, evidence-aware research memory with human review, revision history, and an inspector, allowing assistants to preserve source reports, inferences, assumptions, and evidence across sessions.6 npmMIT
- AlicenseNot gradedqualityAmaintenanceEnables an AI assistant to build and revise an explicit shared world model, then reason, create, narrate, remember, and give writing feedback through inspectable processes, concepts, and cuts. Supports general-purpose modeling, storytelling, agent/user memory, and ideation over one engine and graph while keeping estimates, interpretations, and observations distinct.MIT