This is a structured QA workflow that uses agent-device to systematically explore mobile apps and document bugs with reproducible evidence. You give it an app name or bundle ID, and it launches a testing session that captures screenshots, optional videos, and step-by-step reproduction instructions for every issue it finds. It handles the tedious parts: setting up output directories, managing session state, re-snapshotting after UI changes, and maintaining a structured markdown report with severity categories. The approach is explore-and-document-in-one-pass rather than explore-then-write-up, which means evidence gets captured while context is fresh. Most useful when you need credible bug reports with actual proof rather than vague descriptions.
npx -y skills add callstackincubator/agent-skills --skill dogfood --agent claude-codeInstalls into .claude/skills of the current project.
Select a file.
wshobson/agents
dbt-labs/dbt-agent-skills
github/awesome-copilot