1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
|
# Example configuration with MCP enabled
# This shows how to configure MCP servers with HTTP transport
provider: dmr
model: huggingface.co/jetbrains/mellum2-12b-a2.5b-instruct-gguf-q4_k_m:Q4_K_M
# Yolo mode. `true` (default): the built-in tools run without asking, and each MCP tool
# follows mcp.permissions below. `false`: every tool call waits for your answer —
# [a] once, [s] for the session, [r] reject, [n] never — MCP tools included, even "auto".
yolo: true
# Enable MCP client
mcp:
enabled: true
permissions: ask # Default for MCP tools: ask before executing
discoveryTimeout: 5s # Timeout for discovering tools at startup
servers:
# Example: context7.com MCP server
- name: context7
transport: http
url: https://mcp.context7.com/mcp
timeout: 30s
auth:
type: bearer
tokenEnv: CONTEXT7_TOKEN # Set this environment variable with your token
# Optional: filter which tools to expose
# tools:
# - search
# - get_documentation
# Example: Another HTTP MCP server with API key auth
- name: my-api
transport: http
url: https://api.example.com/mcp
timeout: 15s
auth:
type: api_key
apiKeyEnv: MY_API_KEY
apiKeyHeader: X-API-Key # Optional: custom header name
headers:
X-Custom-Header: custom-value
# Example: Stdio MCP server (NOT YET IMPLEMENTED)
# This will fail with "stdio transport for MCP is not yet implemented"
# - name: local-mcp
# transport: stdio
# command: mcp-server
# args: ["--stdio"]
# env:
# API_KEY: "{{.env.LOCAL_API_KEY}}" # Will be replaced with os.Getenv("LOCAL_API_KEY")
# Rest of the configuration (optional overrides)
maxOutput: 16000
maxTurns: 10
previewLines: 20
# Note: When MCP is enabled, all discovered tools from the servers will be
# available to the model, prefixed with the server name.
# Example: context7_search, context7_get_documentation, etc.
|