> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.rightbrain.ai/v-1/changelog/2026/5/1/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.rightbrain.ai/_mcp/server.

# Skills, Connections, and Agent Versioning

> Manage versioned skills and connections, select active agent revisions, and configure agent fallback models.

## What's New

### Skills

Give your agents reusable, focused abilities with [Skills](/docs/concepts/skills). Browse ready-made Skills from the marketplace or build your own with instructions and reference files, all managed through a new create-and-detail interface. Skills are versioned, so you can iterate safely and pin the exact revision an agent uses. Skills can also declare the tools they depend on, and those dependencies are provisioned onto your agent automatically when the Skill is attached.

### Unified Connections

[Connections](/docs/concepts/connections) brings integrations and MCP servers together in one place. Add native service connectors and Model Context Protocol servers, manage their authentication, and wire them into your agents from a single page.

### Agent Versioning

Changes to an agent’s execution configuration now create a new [revision](/docs/production/revisions) instead of overwriting the live configuration. Pick the active revision from a dropdown in the configure panel, tag revisions for reference, and roll back at any time. In-progress runs keep using the revision they started on.

### Agent Fallback Models

Choose a fallback model for your agents so a run can continue on a second model if the primary one fails. The fallback model name now appears in your run logs when it is used.

## Improvements

* Google integrations gained richer tool metadata for more reliable results
* Permission controls now extend to Skills, Connections, Collections, and agents
* The `url_fetcher` input processor added a direct fetch mode
* More reliable agent streaming, with the session identifier emitted as soon as a run starts
* Skill files render as formatted Markdown and JSON in read-only views
* Renamed the data view to Logs, with clearer OAuth error messages when adding a connection

## Current documentation

[Skills](/docs/concepts/skills) · [Connections](/docs/concepts/connections) · [Revisions](/docs/production/revisions) · [Fallbacks](/docs/production/fallbacks)

> Build reliable AI agents that run inside your existing tools and workflows. Rightbrain developer documentation.