After installing skills (twg skills install --global), give your agent natural language prompts. The agent discovers the right commands, runs them, and synthesizes results.

Simple Prompts

Prompt	What the Agent Runs
"Summarize the work I did in the last 2 weeks"	`twg work query --scope me --since 14d`
"What are my open Jira issues?"	`twg jira issues search --jql "assignee = currentUser() AND statusCategory != Done"`
"Show latest updates on project ATLAS"	`twg projects get ATLAS`
"Who do I collaborate with most?"	`twg collaborators --scope me`
"Find Confluence pages about onboarding"	`twg confluence search query --cql "text ~ 'onboarding'"`
"Create a Jira issue for the login bug"	`twg jira issues create --project-key PROJ --issue-type Bug --summary "Login bug"`
"What's my org structure?"	`twg org-tree --scope me --up-only`
"Show PRs I authored this week"	`twg bitbucket pull-requests query --scope me --since 7d`
"What goals is my team tracking?"	`twg goals --scope team`
"Get full context around PROJ-456"	`twg context jira workitem PROJ-456`

Workflow Prompts

These prompts chain multiple commands together:

Weekly status report:

"Using twg skill, generate a weekly status report for my goals and projects. Include Jira progress, merged PRs, and Confluence updates from this week."

Onboard to a new team:

"Using twg skill, I just joined Platform Engineering. Orient me — team members, org chart, goals, projects, focus areas, recent Confluence docs, current Jira sprint."

Full issue context:

"Using twg skill, explore the full context of PROJ-1234. Show linked PRs, Confluence docs, parent/child issues, and dependencies. Recommend which repo I should clone."

Look someone up:

"Using twg skill, who is jane.doe@company.com? Show her role, org chart, teams, recent work, goals, and how we've collaborated."

Strategic health check:

"Using twg skill, generate a strategic health report for the 'Platform' focus area. Show the tree with position counts, active goals with status, and flag anything off-track."