Arcade MCP

In this catalog it is the Tool Integration option when you need to call authenticated external systems from agents. Keep Arcade MCP on the observe/propose side of the loop: useful signals and drafts, with human or policy authority still required before anything irreversible runs.

Arcade packages authenticated tools as MCP-native capabilities for agent hosts.

1. Define/register tools with auth requirements.
2. MCP client discovers Arcade tools.
3. Agent invokes tools through MCP.
4. Auth’d SaaS/actions execute; results return.

Related: composio · model-context-protocol · topics/33-skill-registry

• Use when: Arcade is an MCP-oriented framework/gateway for custom authenticated agent tools — packaging tool capabilities with auth in an MCP-native way matches the bottleneck and you specifically need to call authenticated external systems from agents.
• Skip when: another pack already covers the need with less ops surface, or Arcade MCP would become a second place people retype truth.
• Prefer instead: compare Composio, Model Context Protocol before locking the cell.

• Ops: confirm license, hosting, and release cadence for Arcade MCP (ArcadeAI/arcade-mcp) before relying on it in a live corridor.
• Integration: bad credentials, scopes, or schema drift produce empty retrievals or runaway tool/model spend without an obvious UI alarm.
• Scope: Tool Integration components support agents; they are not the company ledger or the approval authority.