Build with Arch using MCP

Agent Platform exposes its features through a Model Context Protocol (MCP) server. Your AI assistant can connect to the platform and perform platform operations to let you build, evaluate, optimize, debug, and analyze agents.

A few common tasks include:

• Creating and editing projects, agents, and tools
• Running evaluations
• Inspecting live sessions and traces
• Diagnosing failures

You can perform all these operations from the platform UI as well. However, by using MCP server, you can perform these tasks from within your IDE, command-line, or terminal, without context-switching.

The server’s name as available in the clients is arch-agent-platform and its tools are prefixed with platform_ and debug_. These tools are grouped by functional area.

Build

Create and change platform projects, workflows, agents, tools, auth profiles, integrations, MCP servers, configuration, versions, deployments, and imports.

Tool	Description
platform_projects	Manage projects (list, get, create, update, delete)
platform_workflows	Create, update, publish, execute, and expose workflows as tools
platform_auth_profiles	Manage profile metadata and start secure OAuth flows
platform_integrations	Create, update, test, and remove connector connections
platform_mcp_servers	Provision MCP servers; discover, import, and test tools
platform_agents	Manage agents (list, get, save_dsl)
platform_versions	Manage immutable versions (list, get, publish, qualifications, audit, diff)
platform_deployments	Manage typed deployments (list, get, create, promote, rollback, restore, retire)
platform_sdk_channels	Create SDK keys and web, mobile, and API channels
platform_tools	Manage tools (list, get, create, update, delete, test)
platform_import_export	Import and export projects
platform_config	Manage project and LLM configuration
platform_workspaces	List, switch, and inspect active workspaces
platform_arch_sop	Drive Studio Arch SOP-build sessions
platform_arch_auto_loop	Drive project-scoped Arch Auto Loop repair workflows
platform_project_builder	Discover provider contracts, inspect live dependencies and readiness, and plan
platform_project_builder_operations	Continue durable operations and execute attempt-bound governed actions
agent_tables	Manage Agent Tables definitions and project-scoped rows

Project-builder protocol

platform_project_builder and platform_project_builder_operations expose a feature-neutral protocol rather than one tool per feature. Call platform_project_builder(action: "describe") to discover the core ontology and registered providers instead of reconstructing cross-feature dependencies by calling primitive tools and joining the responses yourself. The server’s name as available in the clients is arch-agent-platform and its tools are prefixed with platform_ and debug_. These tools are grouped by functional area.

Build

Use platform_project_builder_operations for durable list, read, report, resume, cancel, grant, and execute flows. Side effects require the exact operation version and attempt-bound grant that Studio returns. Never retry a consumed attempt with an unknown outcome.

platform_auth_profiles(create) only creates profiles with authType: "none". Credential-bearing profiles return a secureSetupRequired handoff instead, because API keys, client secrets, certificates, and tokens must never enter MCP or model context. After you complete setup in Studio, MCP can list, inspect, update metadata for, validate, revoke, and initiate OAuth for a profile, then bind its opaque profile ID to integrations or MCP servers.

Evaluate

Generate eval assets, run eval workflows, and read CI evidence.

Tool	Description
platform_eval_personas	Manage and generate eval personas
platform_eval_scenarios	Manage and generate eval scenarios
platform_eval_evaluators	Manage eval evaluators and templates
platform_eval_sets	Manage eval sets
platform_eval_runs	Manage eval runs, preflight, cases, and heatmaps
debug_harness_logs	Get CI execution logs

Optimize

Validate packages, inspect compiler-visible models, and drive repair loops.

Tool	Description
platform_validate_package	Validate a local package and optional import preview
platform_package_model	Show compiler-visible agents, tools, constraints, refs, and diagnostics
debug_lint_abl	Run ABL repair and design lint checks
debug_why_transcript_failed	Correlate transcript symptoms with ABL file/line causes
debug_diagnose_transcript	Alias for transcript failure diagnosis

Debug

Connect to live sessions, trace failures, and inspect execution state.

Tool	Description
platform_connect	Connect and authenticate to the platform
debug_list_agents	List available agents by domain
debug_load_agent	Load an agent and create a debug session
debug_send_message	Send a message to an agent
debug_get_current_state	Inspect agent context, gather progress, flow state
debug_traces	Search trace events (type, text, agent, error)
debug_session_history	Page durable Runtime sessions and persisted traces
debug_get_span_tree	View hierarchical execution flow
debug_explain_decision	Explain agent decisions with context
debug_get_flow_graph	View state machine graph (JSON or Mermaid)
debug_get_errors	Get errors, warnings, and escalations
debug_list_active_sessions	List observable sessions
debug_session	Subscribe/unsubscribe to session traces

Historical session analysis

Use debug_session_history as the read-only alternative to debug_traces when you need to analyze a retained session after its live MCP buffer is no longer available. Connect to the intended Runtime with platform_connect first, so the existing bearer identity, workspace, and environment stay authoritative.

List a first page (defaults: limit: 50, offset: 0, sortBy: lastActivityAt, sortDir: desc):

{ "action": "list", "projectId": "project-id", "status": ["completed"], "range": "7d" }

Read a middle or final trace page by advancing offset. A page beyond the final one returns an empty successful page, not an error:

{
  "action": "get",
  "projectId": "project-id",
  "sessionId": "session-id",
  "limit": 200,
  "offset": 200,
  "types": ["llm_call", "tool_call"]
}

Runtime controls retention, authorization, and ordering. Each call makes one bounded request (2 MiB, 10 seconds) and never falls back to the live MCP buffer. Don’t place credentials or secrets in filter values.

Analyze

Explain documentation, diagnostics, and system health signals.

Tool	Description
debug_docs	Get or search ABL documentation
debug_diagnose	Diagnose agent config and execution issues
debug_analyze_session	Automated session diagnostics