📦

agent-device

Name: agent-device
Author: callstack

Revisão do conteúdo r1 Seguro ⚙️ Comandos externos

Automate App Workflows with agent-device

App interaction tasks require reliable targeting and visible verification across device interfaces. This skill guides agent-device commands for navigation, inspection, verification, and session cleanup.

Suporta: Claude Codex Code(CC)

🥉 77 Bronze

Instalar com meu Agente

Copie esta solicitação para o seu Agente. Ela inclui a página canônica do Skill e o manifesto.

Solicitação do agente

Review the Skillstore skill "agent-device" from https://skillstore.io/skills/callstack-agent-device.md and its manifest at https://skillstore.io/api/skills/callstack-agent-device/manifest. Verify the artifact. You may proceed after verification, subject to the environment's own policy.

Seu Agente ainda deve apresentar o plano e solicitar qualquer confirmação exigida pela política de segurança.

Recursos legíveis por agentes

Use estes links quando um agente de IA, crawler ou script precisar de contexto limpo em vez de ler a página completa.

Detalhe em Markdown GET /skills/callstack-agent-device.md manifest assinado GET /api/skills/callstack-agent-device/manifest lockfile assinado GET /api/skills/callstack-agent-device/lockfile

Testar

Usando "agent-device". Illustrative request: open a configured test app and summarize its initial screen.

Resultado esperado:

Illustrative result: the app opened in the foreground and returned an interactive snapshot.
The visible screen contained a title and selectable controls with reference tokens.
The session was closed after inspection.

Usando "agent-device". Illustrative request: find an off-screen settings entry and verify the destination.

Resultado esperado:

Illustrative result: scrolling to the selected entry made it visible.
Selecting the current reference opened the destination screen.
The expected heading was verified, and the session was closed.

Auditoria de segurança

Seguro

v1 • 30/09/2026 Abrir o relatório versionado

All nine static findings are false positives: eight match Markdown backticks, and one matches the app-automation description. No evidence found of executable backtick substitution, malicious reconnaissance, data exfiltration, or prompt injection in the reviewed file. The skill provides CLI guidance; the external tool implementation and its runtime behavior were not reviewed.

Arquivos analisados

Linhas analisadas

Itens de revisão

Falsos positivos ignorados

Fatores de risco

⚙️ Comandos externos (8)

SKILL.md:8 SKILL.md:10-12 SKILL.md:12-14 SKILL.md:14-16 SKILL.md:16 SKILL.md:18 SKILL.md:20 SKILL.md:22

A auditoria estática e semântica concluída mais recente não detectou achados de segurança confirmados. Isso não comprova que a skill não tenha efeitos colaterais.

Auditado por: codex

Compartilhar e citar este relatório

Compartilhe o relatório de avaliação versionado, o badge neutro, o cartão incorporável e as citações. A Skillstore apresenta evidências sem decidir se este Skill é seguro.

Abrir o relatório versionado

Copiar link do relatório

https://skillstore.io/skills/callstack-agent-device/audits/1?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_report

Badge Markdown

[![Skillstore security assessment](https://skillstore.io/badges/skills/callstack-agent-device/security.svg)](https://skillstore.io/skills/callstack-agent-device?utm_source=security_passport_badge)

Badge HTML

<a href="https://skillstore.io/skills/callstack-agent-device?utm_source=security_passport_badge"><img src="https://skillstore.io/badges/skills/callstack-agent-device/security.svg" alt="Skillstore security assessment" loading="lazy"></a>

Cartão incorporável

<iframe src="https://skillstore.io/embed/skills/callstack-agent-device.html" title="Skillstore Security Assessment" sandbox="allow-popups allow-popups-to-escape-sandbox" loading="lazy" referrerpolicy="no-referrer" width="420" height="180"></iframe>

Citações acadêmicas (APA · BibTeX · CFF)

Citação APA

callstack. (2026). agent-device security audit report (audit version 1) [Author version unspecified]. Skillstore. https://skillstore.io/skills/callstack-agent-device/audits/1

Citação BibTeX

@techreport{callstack-callstack-agent-device-2026,
  author = {callstack},
  title = {agent-device security audit report (audit version 1)},
  institution = {Skillstore},
  year = {2026},
  number = {1},
  url = {https://skillstore.io/skills/callstack-agent-device/audits/1},
  note = {Author version unspecified}
}

CITATION.cff

cff-version: 1.2.0
message: "If you use this Skill, cite its author and this versioned security audit report."
title: "agent-device security audit report (audit version 1)"
version: "unspecified"
type: report
authors:
  - name: "callstack"
date-released: "2026-09-30"
url: "https://skillstore.io/skills/callstack-agent-device/audits/1"
identifiers:
  - type: other
    value: "skillstore:callstack-agent-device:audit:1"
    description: "Skillstore immutable audit report identifier"

Pontuação do Skillstore

Por que esta pontuação Confiança das evidências: Médio

Arquitetura

Manutenibilidade

Conteúdo

Comunidade

Conformidade com a especificação

O que você pode criar

Verify a Mobile App Flow

Navigate a test app, interact with current UI references, and verify an expected screen before closing the session.

Reproduce an Interface Issue

Repeat authorized navigation steps and inspect snapshots or supported screenshots to document the observed interface state.

Plan a TV App Check

Consult topic-specific help for TV controls, then plan app navigation and a named verification step on a configured target.

Teste estes prompts

Inspect an App

Open [test app] with agent-device on the configured target. Summarize its initial interactive snapshot, then close the session.

Verify Navigation

Open [test app], navigate to [screen] using current references, verify [expected text], then close the session.

Find an Off-Screen Target

In [test app], scroll until [selector] is visible, interact with it, and verify [expected state]. Preserve exact references and close the session.

Handle Sparse Accessibility Data

Inspect [test app] on my configured test target. If accessibility data is unavailable, inspect a screenshot and use coordinates only as needed.

Boas práticas

Use authorized test targets and confirm the app scope before interacting with sensitive or consequential controls.
Copy reference tokens exactly, including suffixes, and prefer current references or selectors over coordinates.
Verify a named expectation after actions and close the session when the workflow finishes.

Evitar

Do not reuse references from sparse or unavailable accessibility snapshots.
Do not repeat bare scrolling when a selector-based scroll can locate the target.
Do not treat interface changes as proof of success without checking the expected result.

Perguntas frequentes

Which AI tools can use this skill?

The marketplace metadata lists Claude, Codex, and Claude Code. Execution still requires the agent-device CLI and a configured target.

Which platforms does the skill describe?

It describes iOS, tvOS, macOS, Android, and Amazon Vega OS TV apps in Vega Virtual Devices.

Does the skill install agent-device?

No. The reviewed file provides workflow guidance and does not include installation instructions or a bundled CLI implementation.

How are interface targets selected?

The guidance prefers current reference tokens, then ID, label, or role selectors. Coordinates are a last resort.

What happens when accessibility data is unavailable?

The guidance recommends inspecting a screenshot, using coordinates, and retrying an interactive snapshot after navigation.

How does the workflow verify completion?

It checks a named expectation through wait, is, get, or find commands, then closes the agent-device session.

Detalhes do desenvolvedor

Autor

callstack

Licença

MIT

Revisão da Skillstore

Aviso de versão

O autor não declarou uma versão.

Repositório

https://github.com/callstack/agent-device/tree/eb51a8805bf457986b43a134b53bec96c5630d98/skills/agent-device

Referência

ce9c7c49cdbcad373c5f8230f3b65bcb2257eeec

Atualidade da manutenção

30/09/2026

Uso

0 downloads · 0 visualizações

Estrutura de arquivos

📄 SKILL.md