Integrations for OpenClaw, Claude Code, Codex, Gemini, Copilot, and OpenCode.
Connect the coding agents your team already uses, then apply the same rules, permissions, and review evidence across each runtime.
Know where the control plane ends and where each runtime begins.
Compile the same control model into runtime-specific files and config.
Every runtime keeps its own surface. The adapter makes those differences explicit.
Deep filesystem and config mapping with skills, approvals, and manifests.
A similar vocabulary for rules, skills, permissions, and local project settings.
Task-specific skills and local workspace control with a lightweight runtime model.
Coverage for the runtimes users already ask about.
The adapter explains what can be written, what can be observed, and what must stay manual.
Each runtime gets the same product story, but the implementation surface stays honest.
Where the skill folder lives and how it is updated.
What the runtime can allow, deny, or ask for.
What changes can take effect without a restart.
How the control plane learns whether the runtime matched the desired state.
Start with the agent runtime your team already trusts.
Each runtime keeps its own files, permissions, and setup rules. Agents Control makes those differences explicit before rollout.
Govern skills, approvals, manifests, and local workspace state.
Apply shared rules and guidance without flattening runtime-specific behavior.
Keep task skills, local workspace policy, and review evidence visible to the team.
Track adapter readiness while keeping governance language consistent.
One runtime per long-tail page.
Choose the runtime you want to govern first.