al-user-verification

Guides users through manual verification of AL/Business Central feature slices in the Web Client.

4|Updated Feb 22, 2026
One-click install
npx skills add https://github.com/FBakkensen/bc-agentic-dev-tools-marketplace --skill al-user-verification-fbakkensen
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: al-user-verification
Source: https://github.com/FBakkensen/bc-agentic-dev-tools-marketplace/tree/main/plugins/al-agentic-dev/skills/al-user-verification
Command: npx skills add https://github.com/FBakkensen/bc-agentic-dev-tools-marketplace --skill al-user-verification-fbakkensen

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Verifying AL/Business Central features requires a human to walk non-automatable scenarios in the BC Web Client, but coordinating containers, recordings, contract checks, and user observations manually is error-prone. This Skill orchestrates the entire verification cycle so functional outcomes gate the slice while usability feedback becomes tracked findings. ## Core Features & Use Cases - Guided scenario walkthroughs: Presents each non-recorded Journey Example one card at a time with ask-before-reveal questions, so the user's observed values gate the verify task without being led. - Container lifecycle automation: Runs three fresh container spawns per cycle — pre-flight replay of recorded page scripts, contract checks plus the guided walk, and cleanup — isolating verification from prior state. - Failure routing and gating: Flips verify tasks to done or blocked, strips stale review evidence, routes failures to /al-steer with the correct replan trigger, and converts usability observations into candidate task files. - Use Case: After implementing a sales-order slice, run this Skill to replay prior recordings on a fresh container, walk the user through three unrecorded scenarios in the Web Client, capture their verbatim observations, and flip the verify task done before opening the next slice. ## Quick Start Ask the agent to run the al-user-verification skill on the ready-for-verification verify task in the current feature branch's tasks folder.

Frequently Asked Questions about al-user-verification

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I verify a Business Central feature slice with user testing?

Run this Skill on a verify task marked ready-for-verification in the specs tasks folder. It replays recorded scenarios on a fresh container, then guides the user through each non-recorded Journey Example one scenario at a time, recording observed values against expected checks.

What is the difference between Record yes and Record no Journey Examples?

Record yes examples were already walked live while recording page scripts and are only re-confirmed by automated replay in the pre-flight batch. Record no examples cannot be automated, so the user walks them manually in the Web Client and their observations gate the verify task.

Does this Skill work without recorded page scripts?

Yes. A Verification Plan containing only Record no examples, contract examples, or exploration charters is valid. The pre-flight replay batch is skipped and the guided walk or contract checks run directly on a fresh container.

What happens when a functional check fails during verification?

The walk stops immediately at the first functional failure. The Skill records observed versus expected values inline, flips the verify task to blocked, strips the review clean flag, and routes to /al-steer with replan trigger 8 for triage.

Can a verification session be resumed after interruption?

Yes. Completed scenarios stay closed via an inline partial-run record, and re-entry resumes at scenario granularity. The in-flight scenario restarts from its first action because fresh containers erase prior data, while closed scenario verdicts stand.

Why are usability observations not gating the verify task?

Usability findings are subjective testing outcomes, not functional checks, so they never block a slice. Instead they are classified, confirmed with the user, and materialized as candidate task files in the slice for later adjudication.