Settings

Agent-scoped MCP entries for `fabro exec`.

Tool permission level for `fabro exec`.

Model name for `fabro exec`.

LLM provider for `fabro exec`.

Prevent idle sleep on macOS while an exec run is in flight.

Default CLI log level.

Output format for commands that support machine-readable output.

Default output verbosity.

Required for `type = "unix"`; the absolute Unix socket path.

Explicit transport selection.

Required for `type = "http"`; the API base URL.

Check for new Fabro releases during supported CLI commands.

Adapter registry key, such as `"anthropic"`, `"openai"`, `"gemini"`, or `"openai_compatible"`. Required for new providers.

Agent profile used for project memory, CLI/ACP command selection, and native session routing. Override only when a provider needs profile behavior different from its adapter.

Additional provider names accepted by model routing and fallback config.

API-key auth config. Omit the table entirely for providers that need no API key; any `extra_headers` are still attached.

Ordered credential refs. Accepted forms are `vault:<NAME>`, `env:<NAME>`, and `aws_sigv4` (sign requests from the AWS default credential chain — Bedrock). Literal secret strings are rejected.

Primary API-key header policy. Omit when the provider uses a standard bearer token.

Provider API base URL. Required for most custom OpenAI-compatible providers.

Provider-owned billing algorithm for usage estimates. Override for exceptional providers such as local no-billing runtimes.

Human-readable provider name.

Set `false` to disable a provider after lower-precedence layers define it.

Additional headers attached to provider requests. Values are literal text or `{{ secrets.NAME }}` interpolation strings. Put credentials in a secret and reference them with a token, not a bare literal.

Higher-priority ready providers win unqualified model and default selection; ties use canonical provider ID.

Agent profile override for this model. Model overrides take precedence over provider overrides.

Additional model selectors accepted by routing and fallback config. Aliases may repeat across providers, but one selector cannot identify two models within the same provider.

Opaque identifier sent to this provider's API. An explicitly empty value is invalid.

Billing algorithm override for this model — for models whose billing family differs from their provider's (e.g. Claude served through OpenRouter bills Anthropic-style cache reads/writes).

User-facing reasoning effort values Fabro may send for this model. Can be set explicitly for reasoning models whose provider adapter maps effort to a non-native API shape.

Additional speeds beyond implicit `standard`; do not list `standard`.

Cached input/read cost in USD per million tokens.

Input cost in USD per million tokens.

Output cost in USD per million tokens.

Whether this is the provider default model.

Human-readable model name.

Set `false` to disable a model after lower-precedence layers define it.

Estimated output tokens per second for catalog display and planning.

Family label used for catalog display and matching.

Whether prompt cache pricing/usage applies.

Whether the model has reasoning behavior.

Whether requests reason when no `reasoning_effort` is supplied. Set this explicitly for always-reasoning routes that do not expose an effort control, or for effort-capable routes whose provider defaults reasoning off.

Whether the model endpoint supports a native reasoning-effort parameter. `levels` accepts discrete effort levels; `always_adaptive` accepts effort levels with natively always-on adaptive thinking; `none` has no native effort parameter.

Whether the model accepts classic sampling parameters (`temperature`, `top_p`).

Whether the model supports tool calls.

Whether the model accepts image inputs.

Public knowledge cutoff label; TOML dates normalize to `YYYY-MM-DD`.

Maximum context window size in tokens.

Maximum output tokens, if known.

Whether this model should be preferred for provider connectivity probes. Set `false` in a higher-precedence layer to clear an inherited probe marker.

Training data cutoff label.

Allow workflow agents to use Fabro run-management tools.

Agent-scoped MCP server entries, keyed by name.

Command and arguments for `stdio` or `sandbox` transports.

Additional environment variables for process-launching transports.

HTTP headers for `http` transport.

Sandbox port for `sandbox` transport.

Shell script alternative to `command` for process-launching transports. A `stdio` script runs on the host through `sh -c`; a `sandbox` script is evaluated inside the sandbox by non-login Bash.

Max duration for startup and MCP handshake.

Max duration for a single MCP tool call.

MCP transport type.

Remote MCP URL for `http` transport.

Git author email for checkpoint commits.

Git author name for checkpoint commits.

Model-keyed fallback chains. Each value is an ordered list of bare providers, bare model IDs or aliases, or provider-qualified `provider:selector` values. A qualified selector may be a model ID, alias, or provider API ID. Legacy `provider/model` values remain accepted. Each list supports the `...` splice marker at layering time. Fabro selects one chain from the original requested model. A fallback target never activates another model's chain. Model keys stay unresolved in this sparse layer because `fabro validate` is offline and has no server model catalog.

Model name for workflow runs.

Provider name for workflow model selection.

Enable GitHub auto-merge for created pull requests. Implies `draft = false`.

Open created pull requests as drafts.

Automatically create a PR after successful runs.

Merge method to configure for the pull request.