AIMAKE FOR AI CODING AGENTS
Keep your coding agent. Give it a production system of record.
Connect any compatible MCP client to current project context, scoped permissions, governed writes, proof, and durable activity history. Your harness still owns local execution.
External-agent access is included on every paid plan.
REMOTE MCP ENDPOINT
https://api.aimake.dev/mcp
Streamable HTTP
OAuth + PKCE
Project-scoped
THE BOUNDARY
The harness executes. AIMake keeps the work aligned.
Keep the editor, terminal, and model you already trust. AIMake prepares the current why, constraints, and next task before work starts—then records what changed and the evidence behind it.
YOUR HARNESS
Local execution
Edit code, run commands, inspect the product, and work with the repository.
↔
MCP
AIMAKE
Durable control and record
Supply current context, enforce project policy, scope AIMake writes, and retain decisions, proof, and activity.
THE LIVE WORKFLOW
Prepare. Build. Prove. Finish.
Authorize AIMake through your MCP client and select the organization and projects it may reach. After that, every AIMake work session follows one inspectable loop.
01
prepare_work
Prepare with current context
Before code changes, AIMake returns the relevant outcome, task, decisions, Evergreen Notes, repository instructions, and current production policy.
The returned work session is required for AIMake writes.
02
LOCAL HARNESS
Build where you already work
Your coding agent keeps local execution: editing files, running commands, inspecting the application, and working with GitHub.
AIMake governs the work without replacing the harness.
03
add_proof
Return evidence, not a status claim
Attach durable HTTPS proof such as pull requests, test results, logs, dashboards, and demos. Use the web workspace for binary uploads.
Proof requirements stay attached to the outcome they verify.
04
finish_work
Close the loop
Record the factual outcome, verification performed, and concrete follow-ups—even when the work is partial or blocked.
The next person or agent does not have to start cold.
WHY A DURABLE LAYER
Every agent starts from the same truth.
A prompt can explain a workflow once. AIMake keeps the underlying context, permissions, policy, work, and proof current across agents and sessions.
01
Current context, not a copied prompt
The public skill teaches the connection workflow. AIMake supplies private project facts and server-versioned policy when work begins.
02
Scoped and recorded writes
OAuth access is limited to the organizations, projects, and permissions you approve. Tool activity becomes durable history.
03
Proof that stays with the work
Decisions, tasks, delivery evidence, and follow-ups remain connected after the agent session ends.
OPEN BOOTSTRAP, PRIVATE CONTEXT
The skill teaches the connection. AIMake supplies the truth.
The reusable AIMake skill is public, so people and compatible clients can discover how to connect. Your private workspace owns the live project records and the server-enforced production policy.
START WITH ONE PROJECT
Give the next production change the context and proof it deserves.
Connect GitHub, authorize your agent, and keep the work durable from first decision through shipped evidence.
FURTHER READING