Dash0 Raises $110M Series B at $1B Valuation

Last updated: May 31, 2026

Tools & Connectors

Query databases, execute commands, and integrate with external systems.

Agent0 extends its reasoning in two ways: through tools it can always invoke and through connectors to external systems it can access once configured.

Tools

Agent0 has two built-in tools available in every session, with no configuration required.

SQL

Agent0 can execute SQL queries against your connected databases: for joins across signal types, aggregations PromQL cannot express, free-text content searches, and schema exploration.

Bash

Agent0 can execute bounded shell commands inside its sandboxed runtime. Commands run in an isolated environment and cannot affect your production systems directly.

Connectors

Connectors give Agent0 access to external systems outside Dash0. Each connector is configured once per organization in Settings → Integrations.

For each connector, you control:

  • Which users can invoke it — by team or role
  • What it can do — read-only, read and write, or specific tool subsets
  • What it requires — approval before destructive actions, audit logging, rate limits

Configuration is done in Settings → Integrations and applies organization-wide.

GitHub

Agent0 connects to GitHub to:

  • Clone repositories into its sandbox for investigation
  • Read source code, configuration, and commit history
  • Open pull requests with code or instrumentation fixes
  • Add comments and review existing PRs

GitHub is the connector that enables code-level root cause analysis and PR generation.

Linear

Agent0 connects to Linear to:

  • Look up teams, projects, and issues
  • Create new issues from investigation findings
  • Update issue status and comments

MCP (Model Context Protocol)

The Model Context Protocol is the open standard for connecting AI systems to external tools. Agent0 supports any MCP-compatible server as a connector.

Adding an MCP connector

In Settings → Integrations, choose Add connector, select MCP server, and provide the server URL and any required credentials. Agent0 will discover the server's available tools automatically.

Read-only vs destructive tool hints

MCP tools declare whether they are read-only or destructive through standard annotations (readOnlyHint, destructiveHint). Agent0 respects these declarations:

  • Read-only tools can be called without additional approval
  • Destructive tools require explicit configuration before Agent0 will use them
  • Tools that do not declare hints default to requiring approval

What you can connect via MCP

Any tool that exposes an MCP server can be connected to Agent0. For observability workflows, useful connections include knowledge bases (Confluence for runbooks and postmortems), error tracking (Sentry), incident management (PagerDuty), and your own internal APIs. You provide the server URL; Agent0 discovers the available tools automatically.

Note

There is no pre-built connector catalog yet. Each MCP server is added manually. Pre-built onboarding for common tools is on the roadmap.

Connecting your own internal tools

If your team runs an MCP server for internal APIs, Agent0 connects to it the same way as any other MCP server. Any AI tool that supports the protocol, including Agent0, can use it.

Further Reading