agent-first-run-test

Automate zero-shot Android app exploration by driving Clawperator and generating markdown artifacts.

2|Updated Feb 17, 2026
One-click install
npx skills add https://github.com/clawperator/clawperator --skill agent-first-run-test
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: agent-first-run-test
Source: https://github.com/clawperator/clawperator/tree/main/.agents/skills/agent-first-run-test
Command: npx skills add https://github.com/clawperator/clawperator --skill agent-first-run-test

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Automates zero-shot Android app exploration by driving Clawperator when no prewritten skill exists, supporting blind installs, unfamiliar app flows, first-run automation, and documenting the observe-decide-act loop for reusable markdown artifacts.

Core Features & Use Cases

  • Zero-shot exploration: inspect live UI, select stable matchers, execute a single action, and re-observe to progress.
  • Artifact generation: produce structured markdown outputs (execution-log.md, findings.md, and optional skill-draft.md) that capture learnings and potential reusable flows.
  • Reusable groundwork: surface workflows that can be turned into new skills with careful documentation and recovery steps.

Quick Start

Initiate a zero-shot exploration session and let the agent log steps and findings automatically

Frequently Asked Questions about agent-first-run-test

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

FAQPage Schema
How do I automate Android UI testing for an app I have never seen before?

Zero-shot Android automation lets you explore unfamiliar apps by driving an agent through an observe-decide-act loop without prewritten scripts. You inspect the live UI, select stable matchers, execute actions, and re-observe to progress through the interface automatically.

What is zero-shot app exploration and when do I need it?

Zero-shot app exploration is the process of automating an Android interface without prior knowledge of its flows. You need it for blind installs or first-run scenarios where no prewritten skill exists, allowing the agent to document interactions and surface reusable findings.

How do I document an unfamiliar app flow during first-run automation?

You document an unfamiliar app flow by running a zero-shot exploration session that logs the observe-decide-act loop. The agent automatically snapshots the UI, selects matchers, records actions, and generates structured markdown artifacts capturing execution logs and findings.

Can I generate reusable automation scripts from a blind Android app install?

Yes, zero-shot exploration generates a skill-draft.md artifact from a blind install. By documenting the observe-decide-act loop and recovery steps during the session, you surface reusable workflows that can be turned into new permanent automation skills.

Does zero-shot Android automation work without any existing UI test dependencies?

Yes, zero-shot Android automation works with no dependencies. It is designed for scenarios where no prewritten skill exists, relying on the agent to inspect the live UI, select stable matchers, and execute single actions to navigate unfamiliar app flows independently.

What are the limitations of zero-shot exploration for Android UI testing?

Zero-shot exploration relies on selecting stable matchers from a live UI without prior knowledge, which can be fragile for highly dynamic or obfuscated interfaces. It is best suited for documenting unfamiliar flows rather than running high-stakes, repeatable regression tests.