Skill File Overview

What is already in this workspace?

The prepared LAB-11161 workspace includes six agent skills under .github/skills/. A skill is a reusable set of workflow instructions. It helps the agent choose steps and output formats, while Control Hub policy determines which Webex tools are available and you approve side effects.

Skill	When it helps
meeting-follow-up	Turn a completed meeting transcript into a reviewed recap with evidence-backed actions
meeting-quality	Analyze per-participant quality telemetry and separate observations from hypotheses
approval-gate	Preview a Webex write action and request explicit approval
incident-mode	Coordinate a restricted quality incident space and updates
attendance-report	Report attendance without claiming unsupported engagement
multilingual-output	Translate a reviewed output while preserving names, dates, and evidence

Lab 3 covers the manual follow-up workflow and customization of meeting-follow-up. Lab 4 uses the quality server and explores meeting-quality and incident-mode. Lab 5 brings those skills together in one business request.

How skills are arranged

AGENTS.md at the workspace root provides general lab instructions. Each skill directory contains a SKILL.md with a name and description. Some skills link to a references/ directory for the detailed workflow or output format.

LAB-11161/
├── AGENTS.md
├── .github/
│   └── skills/
│       ├── approval-gate/
│       ├── attendance-report/
│       ├── incident-mode/
│       ├── meeting-follow-up/
│       ├── meeting-quality/
│       └── multilingual-output/
├── .vscode/mcp.json
└── src/quality-tool/

Open Configure Skills in Copilot Chat, or type / in the chat input, to see the discovered skills. The agent may load a skill when your request matches its description. Ask it which skill it used and inspect the tool calls; a good final answer alone does not prove the intended skill or data source was used.

What can you customize?

You can change section names, action-item labels, space naming, report formatting, terminology, and troubleshooting checks. Keep transcript evidence, missing-data handling, private-data limits, and approval before Webex writes.

Skills guide behavior

A skill file is not an enforcement boundary. Control Hub can disable Webex tools, and VS Code asks you to review tool calls. If a skill or response proposes an unreviewed write, stop and correct it.