Resource hub
Governance and security
Control permissions, audit trails, safety gates, and approval flows for AI-assisted engineering.
Private AI developmentThe engineering recordLocal model orchestrationAgentic workflowsCodebase intelligence
Platform
Extending an AI coding agent with plugins (2026)
SeaShell plugins contribute new tools to the agent and hook into lifecycle events, so team-specific behavior can be added without waiting for it to ship in the product.
Platform
How SeaShell keeps AI edits safe with permission controls (2026)
SeaShell gates every tool call against allow, ask, and deny rules and shows a unified diff before writing, so automated edits stay auditable and under your control.
Platform
SeaShell and Model Context Protocol (MCP) support (2026)
SeaShell connects to Model Context Protocol servers over stdio, SSE, and HTTP and exposes their tools to the agent, so you can bring your own tools and surfaces.