AgentUI CLI

Use platphormctl for repeatable AgentUI rendering, validation, dry-run, and evidence workflows. Protected actions still require PLATPHORM_API_KEY and explicit operator intent.

CLI Client Guide

site-inspect-agentui

Public-safe

platphormctl site inspect agentui --best-effort

Inspect AgentUI public route, discovery, policy, and health state.

mcp-validate-agentui

Public-safe

platphormctl mcp validate agentui --best-effort

Validate AgentUI MCP JSON-RPC and public introspection.

policy-inspect-agentui

Public-safe

platphormctl policy inspect agentui --best-effort

Inspect AgentUI public/protected policy and robots alignment.

agentui-tools

Public-safe

platphormctl agentui tools

List AgentUI-renderable tool registry entries.

agentui-render-tool

Public-safe

platphormctl agentui render-tool mcp get_health --dry-run

Render the get_health MCP tool as an AgentUI form without executing a protected tool.

agentui-form-schema

Public-safe

platphormctl agentui form-schema mcp get_health --dry-run

Print the deterministic form schema AgentUI derives from an MCP tool schema.

agentui-validate-tool-ui

Public-safe

platphormctl agentui validate-tool-ui mcp get_health --dry-run

Validate required fields, raw JSON fallback, and protected-action labels.

browserops-validate-tool-ui

Protected

platphormctl browserops validate-tool-ui mcp get_health --dry-run

Ask BrowserOps to validate rendered AgentUI tool UI in dry-run mode.

harness-tool-render-check

Public-safe

platphormctl harness run agentui-tool-render-check --dry-run

Run the public AgentUI tool-render harness in dry-run mode.

harness-spec-evals-loop

Protected

platphormctl harness run spec-evals-browserops-loop --dry-run

Run the Spec, Evals, Sandbox, and BrowserOps validation loop in dry-run mode.

protected-workflow-publish

Protected

PLATPHORM_API_KEY=$PLATPHORM_API_KEY platphormctl agentui publish-workflow <workflow-id> --protected --yes

Protected publish shape; requires an operator-provided PLATPHORM_API_KEY value at runtime.