Verify Launch, Rollback, and Maintenance

What you will complete

A complete Business Agent Launch Pack and a final launch, manual, blocked, or stop decision.

Make the operating decision

Install only the validated artifact in the approved scope. Start a fresh task or refresh discovery, invoke the skill explicitly, and run the held-back case against the installed copy. Record the installed path, artifact version, validator output, behavioural verdict, and rollback target.

Write exact invocation, input preparation, expected output, validation, update, disable, and rollback steps. Name where logs and artifacts live without embedding private machine paths in public learner content. Document known limitations, source freshness, ownership, and the event that should trigger maintenance.

Scheduling remains a separate gate. The exact durable prompt must pass manually, then the operator defines cadence, timezone, project, permissions, output, nothing-to-report behaviour, and stopping rule. Do not claim scheduled status until the next run and output destination are verified.

Worked fictional example: Harbourlight Home Services

Harbourlight verifies the installed estimate-exception skill in a fresh task using the held-back export. LAUNCH.md explains how to invoke it, validate the CSV, inspect the brief, update the policy reference, disable discovery, and restore the previous version. Scheduling remains manual.

Harbourlight Home Services and every estimate record in this course are fictional. They demonstrate the method and do not represent a Flowgrammer client, a deployed system, or measured savings.

Complete workbook section 8

Use the evidence available for your own bounded job. Write unknown when the evidence is missing, and record the person or action that can resolve it.

1. Verify runtime discovery and explicit invocation.
2. Run the held-back case against the installed copy.
3. Write invocation, update, disable, rollback, and ownership instructions.
4. Grade the capstone and choose launch, manual, blocked, narrower scope, or stop.

Open the course workbook PDF or read the workbook details.

Critical gate before continuing

• Installed behaviour matches staged behaviour.
• Rollback is executable and names the prior known-good artifact.
• No secret entered instructions or logs.
• Schedule and external integrations remain unclaimed unless tested.

If a gate fails, repair the current section, narrow the scope, leave the route manual, or record a blocked or stop decision. Continuing is not the only successful learner action.

Common failure modes

• Expanding beyond a complete business agent launch pack and a final launch, manual, blocked, or stop decision. before the current artifact can be graded.
• Turning a missing value, unavailable source, or blocked integration into a confident conclusion.
• Treating a prompt instruction as proof that the effective tool or permission boundary works.
• Marking a manual, simulated, or untested route as live.

Check your application