forge-mcp
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| AMIGO_ORG_ID | No | Org to bootstrap and set as session default | |
| AMIGO_API_KEY | No | API key for the bootstrapped org | |
| AMIGO_USER_ID | No | User ID | |
| FORGE_LOG_LEVEL | No | Log verbosity: debug, info, warn, error | |
| AMIGO_API_KEY_ID | No | API key identifier | |
| AMIGO_API_BASE_URL | No | API base URL. Defaults to https://api.amigo.ai |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": true
} |
| resources | {
"listChanged": true
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| forge_set_orgA | Set the active org for this session. Subsequent tool calls will use this org by default. |
| forge_list_orgsA | List all configured orgs with their auth status and show the current session/default org. |
| forge_add_orgA | Add or update credentials for an org. Validates the credentials by signing in. |
| forge_remove_orgB | Remove stored credentials for an org. |
| forge_entity_listB | List all entities of a given type in an org. |
| forge_entity_getA | Get the full details of a specific entity by ID. Returns the latest version. |
| forge_entity_createC | Create a new entity of the given type. Provide the entity data as a JSON string or object. |
| forge_entity_updateB | Update an existing entity by creating a new version. Provide the full version data. |
| forge_entity_deleteA | Delete an entity. This is irreversible. |
| forge_agent_createC | Create a new agent. |
| forge_agent_updateA | Update an agent by creating a new version. The initial version requires: initials, identity, background, behaviors, and communication_patterns. Subsequent versions can include only changed fields. Do NOT include agent_name, greeting, dynamic_behavior_set_ids, user_dimension_ids, or persona_ids. |
| forge_context_graph_createB | Create a new context graph (state machine that defines conversation flow). |
| forge_context_graph_updateB | Update a context graph by creating a new version. All fields are required. The terminal state must have exactly one action. |
| forge_service_createA | Create a new service linking an agent and context graph into a deployable unit. IMPORTANT: Both the agent and context graph must already have at least one version before creating the service. |
| forge_service_updateB | Update an existing service. Only provided fields are changed. |
| forge_tool_createB | Create a new tool (external action the agent can invoke). |
| forge_tool_updateA | Update tool metadata (description and tags). Tool versions are published separately. |
| forge_metric_createC | Create a new metric for evaluating conversations. |
| forge_metric_updateC | Update a metric. Only provided fields are changed. |
| forge_persona_createC | Create a new simulation persona with its initial version. |
| forge_persona_updateA | Update persona metadata (tags). Use forge_entity_update for version changes. |
| forge_scenario_createA | Create a new simulation scenario with its initial version. |
| forge_scenario_updateA | Update scenario metadata (tags). Use forge_entity_update for version changes. |
| forge_dynamic_behavior_set_createB | Create a new dynamic behavior set (runtime instructions injected based on triggers). |
| forge_dynamic_behavior_set_updateB | Update dynamic behavior set metadata. Only provided fields are changed. |
| forge_unit_test_createB | Create a new simulation unit test. |
| forge_unit_test_updateB | Update a unit test. Only provided fields are changed. |
| forge_unit_test_set_createB | Create a new unit test set (collection of unit tests to run together). |
| forge_unit_test_set_updateA | Update a unit test set. Only provided fields are changed. |
| forge_smoke_testA | Run a quick smoke test against a service. Creates a conversation, sends a message, and returns the agent's response. |
| forge_simulateC | Run an automated multi-turn conversation simulation against a service. |
| forge_conversation_insightsB | Get debugging insights for a conversation, including state transitions and triggered behaviors. |
| forge_conversation_evaluateC | Run on-demand metric evaluation against a conversation. |
| forge_version_set_listB | List all version sets for a service, showing pinned versions. |
| forge_version_set_upsertA | Create or update a version set for a service. Use latest=true to pin to the current latest versions. |
| forge_version_set_promoteB | Promote one version set to another (e.g. preview -> release). |
| forge_version_rollbackB | Roll back an entity to a previous version. Lists available versions for selection. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
| dependency-order | Entity dependency order -- create/update entities in this sequence to satisfy references. |
| instructions | Complete Agent Forge instructions for coding agents, including workflows, principles, and best practices. |
TDQS
Scored across 37 tools
Many entity-specific create/update tools (agent, context_graph, metric, etc.) overlap with the generic forge_entity_create/forge_entity_update, and the update semantics are inconsistent: some entity-specific updates create versions, others only edit metadata while pointing to forge_entity_update for version changes. This creates real ambiguity about which tool to call for a given change.
Most tools follow forge_<entity>_<action>, and all use snake_case, but org tools flip to forge_<action>_<entity> (list_orgs, set_org) and a couple are action-only (simulate, smoke_test). The inconsistency is noticeable but not chaotic.
37 tools is well above the 3-15 well-scoped range and even above the 16-25 heavy range; the generic entity_* CRUD layer plus per-entity create/update duplicates much of the surface. A more streamlined set would use either generic CRUD or specialized entity tools, not both.
Core resource lifecycle and testing/evaluation are covered, and entity_delete/get/list fill many gaps. However, unit test sets can be created but not executed, and tool versions are said to be 'published separately' with no corresponding tool, leaving notable dead ends.