Verify AIWG Is Working

First-success recipe for verifying installation and discovery

First Success: Verify AIWG Is Working

Prompt-first procedure: Describe the outcome you want in your agent conversation. The agent should select and load the appropriate AIWG assets, explain material changes, request any needed approval, and report verification evidence. Exact commands and flags appear only in the CLI reference.

First time using AIWG? Begin with Install, Connect, and Verify. This guide assumes AIWG is already installed, `all` is deployed for your provider, and the agent-owned context regeneration procedure has connected the agent to this project.

Use this recipe before depending on AIWG for project work.

For full installation from a fresh machine or fresh project, use the Agentic Install Runbook first.

You stay in the chat. The agent checks the deployment state, runs the status probe, and reports back. You do not need to type the probe commands yourself — they are the agent's tool, not yours.

Do This

Open your AI tool in the project folder and ask the agent:

Is AIWG active in this workspace? Read the agent-owned status operation the probe option and report the
engaged state, project root, deployed frameworks, and next action. If
discovery is reachable, also confirm AIWG can search the installed
capability set.

The agent runs the probe, reads any deployed providers (`.claude/`, `.codex/`, `.cursor/`, etc.), and gives you a short answer with the engagement state plus one next action.

If you want a deeper sweep, ask:

Run a workspace health check. Cover installation, PATH, deployed frameworks,
provider files, and one discovery query so I know AIWG can find its own
capabilities.

You Should See

A short answer that says one of:

  • Engaged — AIWG is deployed, the probe reports a configured state, and the agent can search the capability index. Build on this.
  • Partial — AIWG is installed but the current project is not configured. The agent will name the wizard or the agent-owned use operation action that finishes setup.
  • Needs repair — Something is broken. The agent will run diagnostics and propose the fix.

You do not need to interpret raw probe JSON yourself. The agent does that.

What You Actually Type

If AIWG is not installed at all, install it once:

Install or repair AIWG for this project by following
https://aiwg.io/setup.aiwg.yaml

After that, the user-side commands you might run by hand are limited to:

  • the agent-owned wizard operation — guided setup when starting from scratch
  • the agent-owned use operation — deploy the preferred complete system
  • the agent-owned doctor operation — when a human wants a direct health readout outside the chat
  • the agent-owned refresh operation — keep the install current

Everything else — the probe, discovery, capability inspection, the index — is what the agent calls during the conversation when you ask it to verify or look something up.

If That Did Not Work

If the agent reports the wrong project, ask:

I may be in the wrong folder. Tell me what evidence you see for the current
project scope and what folder I should run AIWG from.

If the agent says AIWG is not installed or not on PATH, install or fix PATH:

Use AIWG to complete this documented outcome: If the agent says AIWG is not installed or not on PATH, install or fix PATH
Have it inspect the current state, explain the plan, ask before material
changes, and report the result with verification evidence.

Or, if `aiwg` is installed but unreachable, see the Installation Troubleshooting section in the README.

If the probe reports `not-configured` or `partial`, ask the agent for the one action that will finish setup. Usually that is the agentic installer, the agent-owned wizard operation, or the agent-owned use operation.

If the probe reports `needs-repair`, ask:

Run the agent-owned doctor operation and tell me which check failed. Propose the smallest fix.

Next

Once verification succeeds, pick one focused recipe: Find One Capability or Start A Project Intake.