Agent Configs
agent_configs
Methods
Create Agent Config -> { id, allowed_tools, created_at, 7 more... }
post/v5/agent_configs
Create Agent Config
List Agent Configs -> CursorPage<{ id, allowed_tools, created_at, 7 more... }>
get/v5/agent_configs
List Agent Configs
List Mcp Tool Names -> Array<"Slack" | "Linear" | "GitHub" | 7 more...>
get/v5/agent_configs/mcp_tools
AllowedTool values that route to MCP servers (vs harness-provided tools). Exposed primarily so the generated SDK includes the McpTool enum type — the frontend asserts its own MCP-tool classifier against this enum to prevent drift.
Get Agent Config -> { id, allowed_tools, created_at, 7 more... }
get/v5/agent_configs/{agent_config_id}
Get Agent Config
path Parameters
agent_config_id: string
Response fields
id: string
allowed_tools: Array<string>
created_at: string
(format: date-time) harness: string
model: string
name: string
system_prompt: string
updated_at: string
(format: date-time) description: string
Optional
object: "agent_config"
(default: "agent_config")Optional
Request example
200Example
Update Agent Config -> { id, allowed_tools, created_at, 7 more... }
patch/v5/agent_configs/{agent_config_id}
Update Agent Config
Delete Agent Config -> { id, deleted, object }
delete/v5/agent_configs/{agent_config_id}
Delete Agent Config
Domain types
AgentConfigCreateResponse = { id, allowed_tools, created_at, 7 more... }
AgentConfigDeleteResponse = { id, deleted, object }
AgentConfigListMcpToolsResponse = Array<"Slack" | "Linear" | "GitHub" | 7 more...>
AgentConfigListResponse = { id, allowed_tools, created_at, 7 more... }
AgentConfigRetrieveResponse = { id, allowed_tools, created_at, 7 more... }
AgentConfigUpdateResponse = { id, allowed_tools, created_at, 7 more... }