arthur-onboard-platform - SKILL.md Agent Skill

name: arthur-onboard-platform description: Onboard an agentic application to the Arthur SaaS Platform (platform.arthur.ai). Guides through authentication, workspace selection, engine deployment, model creation, code instrumentation, trace verification, and eval configuration. allowed-tools: Bash, Read, Write, Edit, Task, Skill version: 1.0.0

Onboard to Arthur Platform

You are guiding the user through the complete Arthur Platform onboarding workflow. Work through each step in order. Be conversational — ask the user before making changes to their code or configuration.

Target repository: The current working directory, unless the user specifies a different path.

Step 0 — Check for skill updates

Invoke the arthur-skills-upgrade skill. It will check all installed arthur-onboard-* and arthur-skills-upgrade skills against GitHub main. If stale skills are found, the user is given three choices:

Yes — upgrade now
Not now — skip this time (will prompt again on the next run)
Skip version — don't prompt again for these specific versions; prompts resume when a newer version is released

If the skill is not installed, skip this step silently.

When upgrades are applied, report the version transition for each updated skill:

"Updated <skill-name>: <old-version> → <new-version>"

If multiple skills were updated, list each one. If everything was already up to date, a brief "All skills up to date" is sufficient.

State File

Persist all state to .arthur-engine.env in the root of the target repository. This file is per-project and should be gitignored.

Before starting: Read the state file:

cat .arthur-engine.env 2>/dev/null || echo "(no state file)"

Parse existing values for ARTHUR_PLATFORM_URL, ARTHUR_ENGINE_URL, ARTHUR_API_KEY, ARTHUR_TASK_ID.

If ARTHUR_ENGINE_URL, ARTHUR_API_KEY, and ARTHUR_TASK_ID all exist, display them and ask:

"Found existing Arthur Platform configuration. Continue with these settings, or start fresh?"

Writing state: Use this pattern to update individual values without clobbering others:

STATE_FILE=".arthur-engine.env"
grep -v '^ARTHUR_PLATFORM_URL=' "$STATE_FILE" 2>/dev/null > /tmp/ae_env_tmp && mv /tmp/ae_env_tmp "$STATE_FILE" || true
echo 'ARTHUR_PLATFORM_URL=https://platform.arthur.ai' >> "$STATE_FILE"

Also ensure the file is gitignored:

grep -qxF '.arthur-engine.env' .gitignore 2>/dev/null || echo '.arthur-engine.env' >> .gitignore

Step 1/13 — Pre-flight Checks + Identify Platform

Check git status in the target repo:

git status --porcelain

Unstaged/untracked changes → warn the user (do NOT block — staged changes are fine)
Not a git repo → note it but continue

Skip Claude Code auth check — the user is already authenticated (they are talking to you right now).

Identify Arthur Platform URL: Ask the user:

"Are you onboarding to the Arthur SaaS Platform at https://platform.arthur.ai? Or do you have a self-hosted Arthur Platform at a different URL?"

Default to https://platform.arthur.ai if the user confirms. Save ARTHUR_PLATFORM_URL to state.

Verify the platform is reachable:

HTTP_STATUS=$(curl -s -o /dev/null -w "%{http_code}" \
  "${ARTHUR_PLATFORM_URL}/api/v1/auth/oidc/.well-known/openid-configuration" 2>/dev/null || echo "000")
echo "PLATFORM_REACHABLE=$HTTP_STATUS"

200 → proceed
anything else → warn the user ("Platform not reachable at "); ask to check the URL or network before continuing

Steps 2–7: Platform Sub-skills

Each step is handled by a dedicated sub-skill. Invoke them in sequence using the Skill tool. Each sub-skill reads its inputs from .arthur-engine.env and writes its outputs back to the same file.

Invoke in order:

Step 2 — arthur-onboard-platform-access Guides service account creation, collects credentials, acquires an OAuth2 token. Establishes ARTHUR_PLATFORM_CLIENT_ID and ARTHUR_PLATFORM_TOKEN in state.
Step 3 — arthur-onboard-platform-workspace Lists or creates a workspace. Establishes ARTHUR_PLATFORM_WORKSPACE_ID and ARTHUR_PLATFORM_WORKSPACE_NAME in state.
Step 4 — arthur-onboard-platform-engine Lists registered engines (data planes) or deploys a new one (Docker Compose, CloudFormation, or Kubernetes). Establishes ARTHUR_PLATFORM_ENGINE_ID and ARTHUR_PLATFORM_ENGINE_URL in state.
Step 5–7 — arthur-onboard-platform-model Gates on application type (Agentic only continues; ML/GenAI models are routed to the platform UI). Creates a Project and Agentic Model on the platform, then retrieves task connection info. Establishes ARTHUR_PLATFORM_PROJECT_ID, ARTHUR_PLATFORM_MODEL_ID, ARTHUR_ENGINE_URL, ARTHUR_API_KEY, and ARTHUR_TASK_ID in state.

Steps 8–13: Reused Sub-skills

After the platform setup is complete, the remaining steps are identical to the OSS onboarding flow. Each sub-skill reads ARTHUR_ENGINE_URL, ARTHUR_API_KEY, and ARTHUR_TASK_ID from state — exactly as set by arthur-onboard-platform-model.

Invoke in order:

Step 8 — arthur-onboard-analyze Analyzes the target repository for language, framework, and existing instrumentation. Writes ARTHUR_DETECTED_LANGUAGE, ARTHUR_DETECTED_FRAMEWORK, ARTHUR_IS_INSTRUMENTED to state.
Step 9 — arthur-onboard-instrument Instruments the application code (Python arthur-sdk, Mastra TypeScript, or OpenInference/OTel).
Step 10 — arthur-onboard-prompts Extracts prompt definitions from the repo and registers them with Arthur Engine.
Step 11 — arthur-onboard-verify Asks the user to run the app, then polls for traces to confirm instrumentation is working.
Step 12 — arthur-onboard-eval-provider Configures an LLM model provider for continuous evals. Writes ARTHUR_EVAL_PROVIDER and ARTHUR_EVAL_MODEL to state.
Step 13 — arthur-onboard-evals Recommends and creates continuous LLM evals for the task.

Sub-skill not found? If a sub-skill is not installed, its step instructions appear in the system's available-skills list. If missing entirely, ask the user to install all arthur-onboard-* and arthur-onboard-platform-* skills alongside this one (see README.md for install commands).

Step 13/13 — Done

After all sub-skills complete, read the final state:

cat .arthur-engine.env 2>/dev/null

Provide a completion summary:

Onboarding complete!

  Arthur Platform:    <ARTHUR_PLATFORM_URL>
  Workspace:          <ARTHUR_PLATFORM_WORKSPACE_NAME> (<ARTHUR_PLATFORM_WORKSPACE_ID>)
  Engine:             <ARTHUR_PLATFORM_ENGINE_ID>
  Engine URL:         <ARTHUR_ENGINE_URL>
  Model:              <ARTHUR_PLATFORM_MODEL_ID>
  Task:               <ARTHUR_TASK_ID>
  Continuous evals:   <N> monitoring your application

Next: Run your application with the Arthur env vars set to start seeing traces and
eval scores in the Arthur Platform UI at <ARTHUR_PLATFORM_URL>.

Note any steps that were skipped or require manual follow-up (e.g., model provider configuration not set, prompts not registered, trace verification pending).