name	fmapi-codingagent-doctor
description	Run comprehensive FMAPI diagnostics — dependencies, configuration, profile, auth, connectivity, and model validation
user_invocable	true

FMAPI Doctor

Run comprehensive diagnostics to identify issues with your Databricks Foundation Model API (FMAPI) configuration.

setup-fmapi-claudecode doctor

Present the output to the user. The doctor runs eight diagnostic categories:
- Dependencies — Checks that jq, databricks, claude, and curl are installed; reports versions
- Environment — Detects OS, WSL version, and headless state
- Configuration — Verifies settings file is valid JSON, all required FMAPI keys are present, and helper script exists and is executable
- Profile — Confirms the Databricks CLI profile exists in ~/.databrickscfg
- Auth — Tests whether the OAuth token is valid
- Connectivity — Tests HTTP reachability to the Databricks serving endpoints API. When AI Gateway v2 is configured, also tests gateway endpoint reachability.
- Models — Validates all four configured model names exist and are ready
- Hooks — Verifies SubagentStart and UserPromptSubmit pre-check hooks are configured and executable
Each check shows PASS, FAIL, WARN, or SKIP with an actionable fix suggestion for failures.
If the command exits with code 1, some checks failed. Guide the user through fixing the reported issues. Common fixes:
- Missing dependencies: install via the suggested command
- Auth failures: suggest /fmapi-codingagent-reauth
- Missing config: suggest /fmapi-codingagent-setup
- Model issues: suggest /fmapi-codingagent-list-models to discover available endpoints

Más de este repositorio

mismo repositorio

Run comprehensive FMAPI diagnostics — dependencies, configuration, profile, auth, connectivity, and model validation

List all serving endpoints available in your Databricks workspace

Re-authenticate Databricks OAuth session for FMAPI

Configure Codex to use Databricks Foundation Model API — interactive or non-interactive setup

Check FMAPI configuration health — OAuth session, workspace, and model settings

Validate that the configured FMAPI model exists and is ready in the workspace

name	fmapi-codingagent-doctor
description	Run comprehensive FMAPI diagnostics — dependencies, configuration, profile, auth, connectivity, and model validation
user_invocable	true

Run comprehensive diagnostics to identify issues with your Databricks Foundation Model API (FMAPI) configuration.

setup-fmapi-claudecode doctor

Present the output to the user. The doctor runs eight diagnostic categories:
- Dependencies — Checks that jq, databricks, claude, and curl are installed; reports versions
- Environment — Detects OS, WSL version, and headless state
- Configuration — Verifies settings file is valid JSON, all required FMAPI keys are present, and helper script exists and is executable
- Profile — Confirms the Databricks CLI profile exists in ~/.databrickscfg
- Auth — Tests whether the OAuth token is valid
- Connectivity — Tests HTTP reachability to the Databricks serving endpoints API. When AI Gateway v2 is configured, also tests gateway endpoint reachability.
- Models — Validates all four configured model names exist and are ready
- Hooks — Verifies SubagentStart and UserPromptSubmit pre-check hooks are configured and executable
Each check shows PASS, FAIL, WARN, or SKIP with an actionable fix suggestion for failures.
If the command exits with code 1, some checks failed. Guide the user through fixing the reported issues. Common fixes:
- Missing dependencies: install via the suggested command
- Auth failures: suggest /fmapi-codingagent-reauth
- Missing config: suggest /fmapi-codingagent-setup
- Model issues: suggest /fmapi-codingagent-list-models to discover available endpoints