Fleet Management

MDM for AI developer tools

Mobile Device Management changed how IT teams deploy software — push an update once, and every device in the fleet receives it. Caliber brings the same model to AI developer tools. Define your approved skills, rules, and MCP configurations centrally, then push-install them to every developer machine in your organization. No more hoping developers manually configure things correctly.

Push-install approved configurations

When your platform team approves a new tool configuration, Caliber deploys it across the fleet. Every developer gets the same approved setup without lifting a finger:

• Skills and rules — push project instructions, coding conventions, and reusable prompt templates to every developer environment
• MCP server configs — deploy approved MCP connections (database access, deployment tools, documentation sources) with correct credentials per team
• Agent definitions — distribute AGENTS.md and CLAUDE.md files that define how AI agents behave in your codebase
• Platform-specific formats — Caliber writes configs in the correct format for each developer's tool (Claude Code, Cursor, Windsurf, Copilot)

Fleet dashboard: deployment status at a glance

The fleet dashboard shows you exactly where every configuration stands across your organization. No guessing, no Slack threads asking “did everyone update their MCP config?”

Deployment policies: auto-deploy vs manual approval

Not every configuration change should roll out the same way. Caliber lets you set deployment policies per tool type so low-risk updates ship automatically while sensitive changes require explicit approval:

• Auto-deploy — rules and skills deploy on next caliber sync. Low risk, high frequency, no approval bottleneck.
• Manual approval — MCP server configs that access production databases or external APIs require team lead sign-off before deployment
• Per-team overrides — platform engineering auto-deploys everything; contractor teams require approval for all tool types

Fleet actions: push, sync, rollback

When configurations need to change, the fleet dashboard gives you three actions to manage the rollout:

• Push update — deploy a new or updated configuration to all developers in a team or across the organization
• Force sync — overwrite local configurations that have drifted from the approved baseline, bringing developers back into compliance
• Rollback — revert to the previous known-good configuration if a push causes issues, applied per-team or fleet-wide

Eliminate configuration drift

Without fleet management, every developer ends up with a slightly different AI tool setup. One person has an outdated MCP config. Another is missing a critical skill. A third has a rule file from three sprints ago. Caliber's fleet management ensures every developer has the same approved configuration — no drift, no surprises, no “works on my machine” for AI tools.

Connected to governance

Fleet management is the enforcement layer for your AI compliance policies. Only tools that pass your governance review get pushed to the fleet. Unapproved tools flagged by shadow AI detection are blocked from deployment. The governance dashboard decides what's approved; fleet management makes sure it actually reaches every developer.

Frequently asked questions

Learn more about team sync for AI configs, or explore how Caliber handles shadow AI detection and AI compliance.