Configuration¶

Configure settings where work executes. A remote client needs the server URL and token; the server owns model, tool, and skill configuration. CLI options override corresponding environment settings.

Model¶

Set LANDING_* variables or save YAML in ~/.landing/config.yml. LANDING_CONFIG selects another file. Empty environment values are ignored; YAML null clears an optional setting.

export LANDING_MODEL="openai:gpt-4.1"
export LANDING_API_KEY="your-provider-api-key"

YAML uses lowercase names without the prefix:

model: openai:gpt-4.1
api_key: your-provider-api-key
completion_args:
  temperature: 0.2

Precedence is environment, Landing YAML, existing Bub YAML, legacy provider-specific environment, then defaults. Corresponding BUB_* aliases remain supported; LANDING_* wins when both are present. Missing YAML values fall back to ~/.bub/config.yml. Provider-specific aliases such as BUB_OPENAI_API_KEY and BUB_OPENAI_API_BASE are lower-priority fallbacks.

Skills¶

Discovery precedence is workspace .agents/skills, explicit roots, configured roots, then ~/.agents/skills. Explicit roots add to configured roots; duplicate skill names select the earlier root. Existing BUB_SKILL_DIRS is a fallback alias. Prepare roots on the executing host; remote clients cannot supply local directories. See Prepare skills.

Mode capabilities¶

YAML modes or JSON LANDING_MODES selects tools and skills independently for each mode:

modes:
  issuer:
    allowed_tools: [spill.read, fs.read, bash, skill, no_update]
    allowed_skills: [issue-triage]
  fixer:
    allowed_tools: [spill.read, fs.read, fs.write, fs.edit, bash, skill]
    allowed_skills: [friendly-python, piglet]
  gatekeeper:
    allowed_tools: [spill.read, fs.read, bash, skill, decide]
    allowed_skills: [landing-review]
  explainer:
    allowed_tools: [spill.read, fs.read, skill]
    allowed_skills: [documentation-writer]

Omitted or null lists are unrestricted; [] disables the collection. Tools accept native names or aliases such as fs.read and fs_read; skill names are case-insensitive and must exist in prepared roots. Include skill for tool-loaded skills, spill.read for oversized output, decide for gate recommendations, and no_update for quiet issuer completion. GitHub publication normally needs bash and authenticated gh; inline reply confirmation also needs confirm_reply.

Per-call SDK selections intersect with mode limits and cannot expand them. Settings apply to CLI, HTTP, SDK, hooks, and CI. Comma commands select modes outside the model loop without adding capabilities. Tool filtering is not a sandbox: shell or delegation tools may expose broader authority.

Execution and service¶

Local workspace defaults to the current directory. Servers register names with serve --workspace NAME=PATH.

GitHub workflows and runner¶

These settings apply to Landing's bundled workflows; a caller's own workflow prepares its environment explicitly.

The runner uses prepared GH_TOKEN or gh login for publication and optional GH_ADMISSION_TOKEN for separate admission. Commit attribution uses Git configuration or GIT_AUTHOR_NAME, GIT_AUTHOR_EMAIL, GIT_COMMITTER_NAME, and GIT_COMMITTER_EMAIL. Selecting a publication token alone does not change commit attribution. See GitHub integration for identity and trust, and Action inputs for invocation.

Container replication¶

Compose forwards only variables listed in compose.yaml, including model, key, endpoint, completion/client options, and model timeout. Add other settings or mounts through an override; exporting them on the host alone does not forward them. See Deployment and Recovery.