Settings Configuration

Fabro loads machine defaults from ~/.fabro/settings.toml. The file is optional. If it does not exist, Fabro falls back to built-in defaults. The CLI and server each read the sections relevant to their own process. On a remote deployment, the CLI machine and the server machine each have their own settings.toml.

Fabro only reads settings.toml. Older cli.toml, user.toml, and server.toml filenames are no longer part of the supported config surface.

File location

The default path is ~/.fabro/settings.toml. Use fabro server start --config /path/to/settings.toml if the server should read a different file.

Schema version

Every Fabro config file must declare its schema version with a top-level _version key:

settings.toml

_version = 1

Files that omit _version are treated as version 1. The legacy top-level version key is no longer accepted and raises a targeted rename hint.

Who reads what

settings.toml uses the same schema as .fabro/project.toml and workflow.toml, but each process only reads the fields it understands. The top-level schema is strictly namespaced — the only allowed domains are [project], [workflow], [run], [cli], [server], and [features].

Scope	Examples
CLI-only	`[cli.target]`, `[cli.auth]`, `[cli.exec]`, `[cli.output]`, `[cli.updates]`, `[cli.logging]`
Shared run defaults	`[run.model]`, `[run.sandbox]`, `[run.checkpoint]`, `[run.inputs]`, `[run.prepare]`, `[run.pull_request]`, `[run.hooks]`, `[run.agent.mcps]`
Server-only	`[server.listen]`, `[server.api]`, `[server.web]`, `[server.auth]`, `[server.storage]`, `[server.artifacts]`, `[server.slatedb]`, `[server.scheduler]`, `[server.logging]`, `[server.integrations]`

[cli.*] and [server.*] stanzas are owner-specific: they are only consumed from ~/.fabro/settings.toml (plus process-local flags and env overrides). The same stanzas in .fabro/project.toml or workflow.toml remain schema-valid but runtime-inert. See Server Configuration for the server-owned sections.

Precedence

Shared layered domains ([project], [workflow], [run], [features]) use this override order:

CLI flags — always win
Environment overrides — Fabro-defined override channels
workflow.toml — per-workflow overrides
.fabro/project.toml — project defaults
~/.fabro/settings.toml — machine defaults
Built-in defaults

Owner-specific domains ([cli.*], [server.*]) use a narrower trust boundary — only CLI flags, env overrides, ~/.fabro/settings.toml, and built-in defaults apply.

Full example

settings.toml

_version = 1

[cli.target]
type = "http"
url = "https://fabro.example.com/api/v1"

[cli.exec]
prevent_idle_sleep = true

[cli.exec.model]
provider = "anthropic"
name = "claude-opus-4-6"

[cli.exec.agent]
permissions = "read-write"

[cli.output]
format = "text"
verbosity = "normal"

[cli.updates]
check = true

[cli.logging]
level = "info"

[run.model]
name = "claude-sonnet-4-5"

[run.git.author]
name = "fabro-bot"
email = "fabro-bot@company.com"

[run.pull_request]
enabled = true

[run.agent.mcps.filesystem]
type = "stdio"
command = ["npx", "-y", "@modelcontextprotocol/server-filesystem", "/workspace"]
startup_timeout = "15s"
tool_timeout = "90s"

[run.agent.mcps.filesystem.env]
NODE_ENV = "production"

[run.agent.mcps.sentry]
type = "http"
url = "https://mcp.sentry.dev/mcp"

[run.agent.mcps.sentry.headers]
Authorization = "Bearer sk-xxx"

All fields are optional. Include only the sections and keys you want to override. A single file can still include both CLI and server sections when you run both processes on one machine, but explicit remote targets do not read remote server state from the local machine.

`[cli.updates]`

Controls whether Fabro runs a daily background check for new releases. The check runs during run, exec, init, and install commands and prints a notice to stderr when a newer version is available.

settings.toml

[cli.updates]
check = true

Key	Value	Description
`check`	`true`	Check for new releases (default)
`check`	`false`	Disable automatic upgrade checks

The --no-upgrade-check CLI flag overrides this for a single invocation. See fabro upgrade for manual upgrades.

`[cli.output]`

Generic CLI output defaults.

settings.toml

[cli.output]
format = "text"
verbosity = "verbose"

Key	Values	Default
`format`	`"text"`, `"json"`	`"text"`
`verbosity`	`"quiet"`, `"normal"`, `"verbose"`	`"normal"`

The -v / --verbose CLI flag always takes effect regardless of this setting.

`[cli.exec]` section

Defaults for fabro exec sessions.

settings.toml

[cli.exec]
prevent_idle_sleep = true

[cli.exec.model]
provider = "anthropic"
name = "claude-opus-4-6"

[cli.exec.agent]
permissions = "read-write"

[cli.exec.model] selects the default LLM for exec:

Key	Description	Values
`provider`	LLM provider	`"anthropic"`, `"openai"`, `"gemini"`, etc.
`name`	Model name	Any model ID from `fabro model list`

[cli.exec.agent] controls agent behavior during exec:

Key	Description	Values	Default
`permissions`	Tool permission level	`"read-only"`, `"read-write"`, `"full"`	`"read-write"`

Permission levels

read-only — auto-approves read tools (read_file, grep, glob, list_dir) and subagent tools
read-write — adds write tools (write_file, edit_file, apply_patch)
full — allows all tools including shell commands

Tools outside the permission level are interactively prompted (if a TTY is present) or denied (with --auto-approve).

`[run.model]` section

Defaults for workflow model selection in commands like fabro run and fabro preflight.

settings.toml

[run.model]
provider = "anthropic"
name = "claude-sonnet-4-5"
fallbacks = ["openai", "gpt-5.4", "gemini/gemini-flash"]

Key	Description	Values	Default
`name`	Model name	Any model ID from `fabro model list`	Per provider
`provider`	Provider name	`"anthropic"`, `"openai"`, `"gemini"`, etc.	Auto-inferred from model/catalog
`fallbacks`	Ordered list of fallback model references	bare provider, bare alias, or `provider/model`	`[]`

Use [cli.exec.model] to configure provider and model for fabro exec. Use [run.model] for workflow-oriented defaults.

`[cli.logging]` section

Configure the default CLI log level. Precedence: FABRO_LOG env var > --debug flag > [cli.logging].level > "info".

settings.toml

[cli.logging]
level = "info"

Key	Values	Default
`level`	`"error"`, `"warn"`, `"info"`, `"debug"`, `"trace"`	`"info"`

Server-side logging is a separate namespace at [server.logging].

`[run.git.author]`

Customize the git author identity used for checkpoint commits.

settings.toml

[run.git.author]
name = "fabro-bot"
email = "fabro-bot@company.com"

Key	Description	Default
`name`	Git author name	`"fabro"`
`email`	Git author email	`"fabro@local"`

`[cli.target]` section

Connection info for commands that target a remote Fabro server.

settings.toml

[cli.target]
type = "http"
url = "https://fabro.example.com/api/v1"

Key	Description
`type`	`"http"` or `"unix"` — explicit transport selection
`url`	Required for `type = "http"` — the API base URL
`path`	Required for `type = "unix"` — the absolute Unix socket path

fabro model uses [cli.target] by default when no explicit --storage-dir is passed. An explicit --server flag overrides the configured target:

fabro model list --server https://fabro.example.com/api/v1

An explicit http(s)://... target is always remote-by-contract. Fabro does not derive auth for that target from a local storage dir, an active local daemon record, or ~/.fabro/dev-token. Use CLI OAuth (fabro auth login --server ...) or an explicit FABRO_DEV_TOKEN when you need remote auth. fabro auth login only works with type = "http" targets. Unix-socket targets use the local dev-token flow instead of browser OAuth. Plain http://... targets are supported for local or trusted deployments; operators remain responsible for providing HTTPS anywhere real credentials cross an untrusted network. fabro exec does not automatically use [cli.target]. It only routes model traffic through a Fabro server when you pass --server for that invocation.

`[run.pull_request]`

Enable auto-PR globally so workflows open a GitHub pull request on successful completion.

settings.toml

[run.pull_request]
enabled = true

Key	Description	Default
`enabled`	Automatically create a PR after successful runs	`false`
`draft`	Open the PR as a draft	`true`
`auto_merge`	Enable GitHub auto-merge on the created PR (implies `draft = false`)	`false`
`merge_strategy`	One of `"squash"`, `"merge"`, `"rebase"`	`"squash"`

Precedence: workflow.toml > .fabro/project.toml > ~/.fabro/settings.toml > built-in default (false).

`[run.agent.mcps]` section

Configure MCP servers to connect to during agent-driven runs. Each server is a named TOML table under [run.agent.mcps]. For fabro exec-only MCPs, use [cli.exec.agent.mcps.*] with the same shape.

Stdio transport

Spawn a local process and communicate over stdin/stdout:

settings.toml

[run.agent.mcps.filesystem]
type = "stdio"
command = ["npx", "-y", "@modelcontextprotocol/server-filesystem", "/workspace"]
startup_timeout = "15s"
tool_timeout = "90s"

[run.agent.mcps.filesystem.env]
NODE_ENV = "production"

Key	Description	Default
`type`	Must be `"stdio"`	—
`command`	Array: executable + arguments	—
`env`	Additional environment variables for the child process	`{}`
`startup_timeout`	Max duration for the MCP handshake (e.g. `"10s"`, `"30s"`)	`"10s"`
`tool_timeout`	Max duration for a single tool call (e.g. `"60s"`, `"2m"`)	`"60s"`

HTTP transport

Connect to a remote MCP server over Streamable HTTP:

settings.toml

[run.agent.mcps.sentry]
type = "http"
url = "https://mcp.sentry.dev/mcp"

[run.agent.mcps.sentry.headers]
Authorization = "Bearer sk-xxx"

Key	Description	Default
`type`	Must be `"http"`	—
`url`	The MCP server endpoint URL	—
`headers`	Optional HTTP headers (for example, for authentication)	`{}`
`startup_timeout`	Max duration for the MCP handshake	`"10s"`
`tool_timeout`	Max duration for a single tool call	`"60s"`

Sandbox transport

Run an MCP server inside the workflow’s sandbox and connect via preview URL. Only available with remote sandbox providers (Daytona) that support port previews. Typically configured in workflow.toml rather than settings.toml:

workflow.toml

[run.agent.mcps.playwright]
type = "sandbox"
command = ["npx", "@playwright/mcp@latest", "--port", "3100", "--headless"]
port = 3100
startup_timeout = "60s"
tool_timeout = "2m"

Key	Description	Default
`type`	Must be `"sandbox"`	—
`command`	Array: the command to run inside the sandbox	—
`port`	Port the server listens on inside the sandbox	—
`env`	Additional environment variables for the server process	`{}`
`startup_timeout`	Max duration for startup + MCP handshake	`"10s"`
`tool_timeout`	Max duration for a single tool call	`"60s"`

See MCP — Sandbox transport for how Fabro launches and connects to sandbox MCP servers.

Documentation Index

​File location

​Schema version

​Who reads what

​Precedence

​Full example

​[cli.updates]

​[cli.output]

​[cli.exec] section

​Permission levels

​[run.model] section

​[cli.logging] section

​[run.git.author]

​[cli.target] section

​[run.pull_request]

​[run.agent.mcps] section

​Stdio transport

​HTTP transport

​Sandbox transport

File location

Schema version

Who reads what

Precedence

Full example

`[cli.updates]`

`[cli.output]`

`[cli.exec]` section

Permission levels

`[run.model]` section

`[cli.logging]` section

`[run.git.author]`

`[cli.target]` section

`[run.pull_request]`

`[run.agent.mcps]` section

Stdio transport

HTTP transport

Sandbox transport