📦

agent-browser

Name: agent-browser
Author: tul-sh

Revisión del contenido r1 Riesgo alto ⚙️ Comandos externos🌐 Acceso a red📁 Acceso al sistema de archivos

Automatizar flujos de trabajo del navegador

Las tareas manuales del navegador son lentas de repetir y difíciles de verificar. Esta skill guía a Claude, Codex y Claude Code a través de flujos de trabajo con scripts para navegación web, captura y pruebas.

Compatible con: Claude Codex Code(CC)

⚠️ 38 Deficiente

Detente y solicita confirmación antes de instalar.

Revisa el plan y obtén el consentimiento explícito del usuario antes de modificar archivos.

Instalar con mi Agente

Copia esta solicitud en tu Agente. Incluye la página canónica del Skill y el manifiesto.

Solicitud de agente

Review the Skillstore skill "agent-browser" from https://skillstore.io/skills/tul-sh-agent-browser.md and its manifest at https://skillstore.io/api/skills/tul-sh-agent-browser/manifest. Verify the artifact. Stop and obtain explicit user consent before installing or changing files.

Tu Agente debe seguir mostrando su plan y solicitar cualquier confirmación exigida por la política de seguridad.

Recursos legibles por agentes

Usa estos enlaces cuando un agente de IA, crawler o script necesite contexto limpio en vez de leer la página completa.

Detalle Markdown GET /skills/tul-sh-agent-browser.md Manifest firmado GET /api/skills/tul-sh-agent-browser/manifest Lockfile firmado GET /api/skills/tul-sh-agent-browser/lockfile

Pruébalo

Usando "agent-browser". Abrir una página de producto y resumir el estado del navegador.

Resultado esperado:

La página se cargó correctamente. El informe incluye el título, la URL actual, los botones visibles, los campos de formulario y una referencia de screenshot.

Usando "agent-browser". Ejecutar una prueba de envío de formulario con datos de muestra.

Resultado esperado:

El flujo de trabajo identificó los campos, rellenó los valores de muestra, hizo clic en enviar e informó si apareció un mensaje de éxito.

Usando "agent-browser". Grabar un bug de navegación sin credenciales.

Resultado esperado:

La sesión produjo una grabación corta, un screenshot final y notas sobre el objetivo de clic fallido.

Auditoría de seguridad

Riesgo alto

v4 • 7/7/2026 Abrir el informe versionado

The skill is a legitimate browser automation wrapper, but it carries high-risk capabilities because it drives a third-party browser service through shell commands. Static path traversal, Markdown backticks, placeholder URLs, /dev/null redirection, and reconnaissance matches were mostly false positives, while screenshots, video, proxy routing, file upload, JavaScript execution, and authenticated-session handling are confirmed risks.

Archivos escaneados

2,107

Líneas analizadas

Elementos de revisión

Falsos positivos ignorados

Preocupaciones de seguridad confirmadas (6)

Alto

SKILL.md:3

Screen capture upload

description: "Browser automation for AI agents via inference.sh. Navigate web pages, interact with e

The skill description advertises screenshots, video recording, JavaScript execution, and file upload. Those capabilities can capture or move sensitive page and local file data through the browser service.

Alto

SKILL.md:3-4 references/commands.md:51-60 references/commands.md:152-185

Third-Party Browser Session Handles Sensitive Page Data

The skill grants Bash access to infsh and describes browser automation through inference.sh. Page snapshots, screenshots, and videos can pass through that external browser service.

The documentation directly states inference.sh usage and shows returned screenshots from open, screenshot, and execute functions. This confirms third-party handling of browser output.

Alto

references/authentication.md:31-38 references/authentication.md:120-144 references/authentication.md:209-224

Credential, 2FA, and Cookie Handling Workflows

Authentication guidance includes entering passwords, generating TOTP codes, reusing authenticated sessions, and extracting cookies. These workflows can expose account access if used without explicit consent and secure handling.

The cited sections explicitly automate password entry, TOTP generation, and cookie extraction. These are high-impact authentication materials, not generic browser actions.

Alto

SKILL.md:156-165 references/commands.md:95-98 references/commands.md:101-113

Local File Upload Capability

The interact action supports uploading local file paths to web pages. An agent using this without careful approval could disclose local documents to a target site.

The documentation lists file_paths for upload and provides a file upload example. The capability is explicit and directly affects local file confidentiality.

Alto

references/proxy-support.md:85-128

Proxy Rotation for Rate Limit Avoidance

The proxy guide includes a section titled rate limit avoidance and shows rotating proxies while scraping pages. This creates business-logic abuse risk beyond basic proxy configuration.

The section title and example are explicit about proxy rotation for scraping. This is a clear intent-level abuse risk, not merely a hardcoded URL pattern.

Alto

SKILL.md:180-189 references/commands.md:162-185 templates/capture-workflow.sh:103-117

Arbitrary JavaScript Execution in Page Context

The execute function runs custom JavaScript on the current page and returns results with screenshots. In authenticated sessions, this can read page content and session-accessible data.

The function is documented as running arbitrary JavaScript, and templates use it to extract page text and links. This confirms page-context code execution as an intended capability.

Elementos de revisión de capacidades (60)

Estas son capacidades locales reales que pueden esperarse para esta habilidad, por lo que requieren revisión, pero no se cuentan como comportamiento malicioso confirmado.

Alto

references/authentication.md:215-217

Shell command substitution

RESULT=$(infsh app run agent-browser --function execute --session $SESSION --input '{

This line runs JavaScript through the browser service inside an authenticated session. In this context it is used for cookie or session data extraction, which is sensitive.

Alto

references/authentication.md:218

Shell command substitution

COOKIES=$(echo $RESULT | jq -r '.result')

This line stores cookies extracted from browser JavaScript output. Handling session cookies is sensitive and can expose authenticated access if logged or reused.

Alto

references/authentication.md:222-223

Shell command substitution

RESULT=$(infsh app run agent-browser --function execute --session $SESSION --input '{

This line runs JavaScript through the browser service inside an authenticated session. In this context it is used for cookie or session data extraction, which is sensitive.

Alto

references/proxy-support.md:152

SOCKS proxy

"proxy_url": "socks5://privacy-proxy.example.com:1080"

The documentation explicitly supports SOCKS5 proxy routing for privacy and anonymity. This can bypass normal network attribution and monitoring, so the proxy capability is a real risk.

Alto

references/proxy-support.md:165

SOCKS proxy

### SOCKS5 Proxy

The documentation explicitly supports SOCKS5 proxy routing for privacy and anonymity. This can bypass normal network attribution and monitoring, so the proxy capability is a real risk.

Alto

references/proxy-support.md:168

SOCKS proxy

{"proxy_url": "socks5://proxy.example.com:1080"}

The documentation explicitly supports SOCKS5 proxy routing for privacy and anonymity. This can bypass normal network attribution and monitoring, so the proxy capability is a real risk.

Alto

references/video-recording.md:23-26

Shell command substitution

SESSION=$(infsh app run agent-browser --function open --session new --input '{

This line controls a browser session with video or screenshot capture in the surrounding workflow. The external command can return recordings or page images containing sensitive data.

Alto

references/video-recording.md:38

Shell command substitution

RESULT=$(infsh app run agent-browser --function close --session $SESSION --input '{}')

This line controls a browser session with video or screenshot capture in the surrounding workflow. The external command can return recordings or page images containing sensitive data.

Alto

references/video-recording.md:48-52

Shell command substitution

SESSION=$(infsh app run agent-browser --function open --session new --input '{

This line controls a browser session with video or screenshot capture in the surrounding workflow. The external command can return recordings or page images containing sensitive data.

Alto

references/video-recording.md:87-90

Shell command substitution

SESSION=$(infsh app run agent-browser --function open --session new --input '{

This line controls a browser session with video or screenshot capture in the surrounding workflow. The external command can return recordings or page images containing sensitive data.

Alto

references/video-recording.md:93-95

Shell command substitution

RESULT=$(infsh app run agent-browser --function interact --session $SESSION --input '{

This line controls a browser session with video or screenshot capture in the surrounding workflow. The external command can return recordings or page images containing sensitive data.

Alto

references/video-recording.md:103

Shell command substitution

CLOSE_RESULT=$(infsh app run agent-browser --function close --session $SESSION --input '{}')

This line controls a browser session with video or screenshot capture in the surrounding workflow. The external command can return recordings or page images containing sensitive data.

Alto

references/video-recording.md:119-124

Shell command substitution

SESSION=$(infsh app run agent-browser --function open --session new --input '{

This line controls a browser session with video or screenshot capture in the surrounding workflow. The external command can return recordings or page images containing sensitive data.

Alto

references/video-recording.md:156

Shell command substitution

RESULT=$(infsh app run agent-browser --function close --session $SESSION --input '{}')

This line controls a browser session with video or screenshot capture in the surrounding workflow. The external command can return recordings or page images containing sensitive data.

Alto

references/video-recording.md:168-171

Shell command substitution

SESSION=$(infsh app run agent-browser --function open --session new --input '{

This line controls a browser session with video or screenshot capture in the surrounding workflow. The external command can return recordings or page images containing sensitive data.

Alto

references/video-recording.md:178

Shell command substitution

CLOSE_RESULT=$(infsh app run agent-browser --function close --session $SESSION --input '{}')

This line controls a browser session with video or screenshot capture in the surrounding workflow. The external command can return recordings or page images containing sensitive data.

Alto

references/video-recording.md:197-200

Shell command substitution

SESSION=$(infsh app run agent-browser --function open --session new --input '{

This line controls a browser session with video or screenshot capture in the surrounding workflow. The external command can return recordings or page images containing sensitive data.

Alto

references/video-recording.md:206

Shell command substitution

RESULT=$(infsh app run agent-browser --function close --session $SESSION --input '{}')

This line controls a browser session with video or screenshot capture in the surrounding workflow. The external command can return recordings or page images containing sensitive data.

Alto

SKILL.md:116-120

Shell command substitution

SESSION=$(infsh app run agent-browser --function open --session new --input '{

This line controls a browser session with video or screenshot capture in the surrounding workflow. The external command can return recordings or page images containing sensitive data.

Alto

SKILL.md:248-251

Shell command substitution

SESSION=$(infsh app run agent-browser --function open --session new --input '{

This line controls a browser session with video or screenshot capture in the surrounding workflow. The external command can return recordings or page images containing sensitive data.

Alto

SKILL.md:259

Shell command substitution

RESULT=$(infsh app run agent-browser --function close --session $SESSION --input '{}')

This line controls a browser session with video or screenshot capture in the surrounding workflow. The external command can return recordings or page images containing sensitive data.

Medio

references/authentication.md:24-26

Shell command substitution

SESSION=$(infsh app run agent-browser --function open --session new --input '{

The line invokes the inference.sh CLI to control a browser session. This is intentional, but it executes an external command that can transmit page state, screenshots, or entered data.

Medio

references/authentication.md:51

Shell command substitution

RESULT=$(infsh app run agent-browser --function snapshot --session $SESSION --input '{}')

The line invokes the inference.sh CLI to control a browser session. This is intentional, but it executes an external command that can transmit page state, screenshots, or entered data.

Medio

references/authentication.md:70-72

Shell command substitution

SESSION=$(infsh app run agent-browser --function open --session new --input '{

The line invokes the inference.sh CLI to control a browser session. This is intentional, but it executes an external command that can transmit page state, screenshots, or entered data.

Medio

references/authentication.md:80

Shell command substitution

RESULT=$(infsh app run agent-browser --function snapshot --session $SESSION --input '{}')

The line invokes the inference.sh CLI to control a browser session. This is intentional, but it executes an external command that can transmit page state, screenshots, or entered data.

Medio

references/authentication.md:97

Shell command substitution

RESULT=$(infsh app run agent-browser --function snapshot --session $SESSION --input '{}')

The line invokes the inference.sh CLI to control a browser session. This is intentional, but it executes an external command that can transmit page state, screenshots, or entered data.

Medio

references/authentication.md:115

Shell command substitution

RESULT=$(infsh app run agent-browser --function snapshot --session $SESSION --input '{}')

The line invokes the inference.sh CLI to control a browser session. This is intentional, but it executes an external command that can transmit page state, screenshots, or entered data.

Medio

references/authentication.md:128

Shell command substitution

RESULT=$(infsh app run agent-browser --function snapshot --session $SESSION --input '{}')

The line invokes the inference.sh CLI to control a browser session. This is intentional, but it executes an external command that can transmit page state, screenshots, or entered data.

Medio

references/authentication.md:133

Shell command substitution

TOTP_CODE=$(oathtool --totp -b "$TOTP_SECRET")

The command generates a TOTP code from a secret in a shell workflow. It handles authentication material and should require explicit user approval.

Medio

references/authentication.md:153-156

Shell command substitution

SESSION=$(infsh app run agent-browser --function open --session new --input '{

The line invokes the inference.sh CLI to control a browser session. This is intentional, but it executes an external command that can transmit page state, screenshots, or entered data.

Medio

references/authentication.md:179-181

Shell command substitution

SESSION=$(infsh app run agent-browser --function open --session new --input '{

The line invokes the inference.sh CLI to control a browser session. This is intentional, but it executes an external command that can transmit page state, screenshots, or entered data.

Medio

references/authentication.md:202

Shell command substitution

SESSION=$(login)

The command substitution runs a login helper that creates an authenticated browser session. That session can carry cookies and permissions into later automation.

Medio

references/authentication.md:243

Shell command substitution

export PASSWORD=$(cat /path/to/secure/password)

The command reads a local password file into an environment variable. Even as a best-practice example, it handles local secrets and can leak if logged or reused carelessly.

Medio

references/authentication.md:246

Shell command substitution

export PASSWORD=$(vault read -field=password secret/app)

The command reads a password from a secrets manager into the shell environment. This is legitimate secret handling but still a sensitive operation.

Medio

references/proxy-support.md:22-25

Shell command substitution

SESSION=$(infsh app run agent-browser --function open --session new --input '{

The line invokes the inference.sh CLI to control a browser session. This is intentional, but it executes an external command that can transmit page state, screenshots, or entered data.

Medio

references/proxy-support.md:35-40

Shell command substitution

SESSION=$(infsh app run agent-browser --function open --session new --input '{

The line invokes the inference.sh CLI to control a browser session. This is intentional, but it executes an external command that can transmit page state, screenshots, or entered data.

Medio

references/proxy-support.md:65-68

Shell command substitution

SESSION=$(infsh app run agent-browser --function open --session new --input '{

The line invokes the inference.sh CLI to control a browser session. This is intentional, but it executes an external command that can transmit page state, screenshots, or entered data.

Medio

references/proxy-support.md:76

Shell command substitution

RESULT=$(infsh app run agent-browser --function snapshot --session $SESSION --input '{}')

The line invokes the inference.sh CLI to control a browser session. This is intentional, but it executes an external command that can transmit page state, screenshots, or entered data.

Medio

references/proxy-support.md:113-116

Shell command substitution

SESSION=$(infsh app run agent-browser --function open --session new --input '{

The line invokes the inference.sh CLI to control a browser session. This is intentional, but it executes an external command that can transmit page state, screenshots, or entered data.

Medio

references/proxy-support.md:119-121

Shell command substitution

RESULT=$(infsh app run agent-browser --function execute --session $SESSION --input '{

The line invokes the inference.sh CLI to control a browser session. This is intentional, but it executes an external command that can transmit page state, screenshots, or entered data.

Medio

references/proxy-support.md:137-142

Shell command substitution

SESSION=$(infsh app run agent-browser --function open --session new --input '{

The line invokes the inference.sh CLI to control a browser session. This is intentional, but it executes an external command that can transmit page state, screenshots, or entered data.

Medio

references/proxy-support.md:150-153

Shell command substitution

SESSION=$(infsh app run agent-browser --function open --session new --input '{

The line invokes the inference.sh CLI to control a browser session. This is intentional, but it executes an external command that can transmit page state, screenshots, or entered data.

Medio

references/proxy-support.md:186-189

Shell command substitution

SESSION=$(infsh app run agent-browser --function open --session new --input '{

The line invokes the inference.sh CLI to control a browser session. This is intentional, but it executes an external command that can transmit page state, screenshots, or entered data.

Medio

references/proxy-support.md:192-194

Shell command substitution

RESULT=$(infsh app run agent-browser --function execute --session $SESSION --input '{

The line invokes the inference.sh CLI to control a browser session. This is intentional, but it executes an external command that can transmit page state, screenshots, or entered data.

Medio

references/proxy-support.md:267-270

Shell command substitution

SESSION=$(infsh app run agent-browser --function open --session new --input '{

The line invokes the inference.sh CLI to control a browser session. This is intentional, but it executes an external command that can transmit page state, screenshots, or entered data.

Medio

references/snapshot-refs.md:124-126

Shell command substitution

RESULT=$(infsh app run agent-browser --function interact --session $SESSION --input '{

The line invokes the inference.sh CLI to control a browser session. This is intentional, but it executes an external command that can transmit page state, screenshots, or entered data.

Medio

references/snapshot-refs.md:136-138

Shell command substitution

RESULT=$(infsh app run agent-browser --function interact --session $SESSION --input '{

The line invokes the inference.sh CLI to control a browser session. This is intentional, but it executes an external command that can transmit page state, screenshots, or entered data.

Medio

SKILL.md:36-38

Shell command substitution

RESULT=$(infsh app run agent-browser --function open --session new --input '{

The line invokes the inference.sh CLI to control a browser session. This is intentional, but it executes an external command that can transmit page state, screenshots, or entered data.

Medio

SKILL.md:215-217

Shell command substitution

SESSION=$(infsh app run agent-browser --function open --session new --input '{

The line invokes the inference.sh CLI to control a browser session. This is intentional, but it executes an external command that can transmit page state, screenshots, or entered data.

Medio

SKILL.md:233-235

Shell command substitution

SESSION=$(infsh app run agent-browser --function open --session new --input '{

The line invokes the inference.sh CLI to control a browser session. This is intentional, but it executes an external command that can transmit page state, screenshots, or entered data.

Medio

templates/authenticated-session.sh:40-42

Shell command substitution

RESULT=$(infsh app run agent-browser --function open --session new --input '{

The line invokes the inference.sh CLI to control a browser session. This is intentional, but it executes an external command that can transmit page state, screenshots, or entered data.

Medio

templates/authenticated-session.sh:1

Unix shell invocation

#!/bin/bash

This is an executable shell template, not just documentation. Running it invokes the inference.sh CLI with user-provided URLs and browser sessions.

Medio

templates/capture-workflow.sh:28

Shell command substitution

CLOSE_RESULT=$(infsh app run agent-browser --function close --session $SESSION_ID --input '{}' 2>/de

The executable template calls inference.sh to open pages, capture screenshots, or execute page JavaScript. It is useful but sends browser state and extracted content through an external service.

Medio

templates/capture-workflow.sh:54-59

Shell command substitution

RESULT=$(infsh app run agent-browser --function open --session new --input '{

The executable template calls inference.sh to open pages, capture screenshots, or execute page JavaScript. It is useful but sends browser state and extracted content through an external service.

Medio

templates/capture-workflow.sh:86

Shell command substitution

RESULT=$(infsh app run agent-browser --function snapshot --session $SESSION_ID --input '{}')

The executable template calls inference.sh to open pages, capture screenshots, or execute page JavaScript. It is useful but sends browser state and extracted content through an external service.

Medio

templates/capture-workflow.sh:105-107

Shell command substitution

RESULT=$(infsh app run agent-browser --function execute --session $SESSION_ID --input '{

The executable template calls inference.sh to open pages, capture screenshots, or execute page JavaScript. It is useful but sends browser state and extracted content through an external service.

Medio

templates/capture-workflow.sh:114-115

Shell command substitution

RESULT=$(infsh app run agent-browser --function execute --session $SESSION_ID --input '{

The executable template calls inference.sh to open pages, capture screenshots, or execute page JavaScript. It is useful but sends browser state and extracted content through an external service.

Medio

templates/capture-workflow.sh:1

Unix shell invocation

#!/bin/bash

This is an executable shell template, not just documentation. Running it invokes the inference.sh CLI with user-provided URLs and browser sessions.

Medio

templates/form-automation.sh:30-32

Shell command substitution

RESULT=$(infsh app run agent-browser --function open --session new --input '{

The line invokes the inference.sh CLI to control a browser session. This is intentional, but it executes an external command that can transmit page state, screenshots, or entered data.

Medio

templates/form-automation.sh:1

Unix shell invocation

#!/bin/bash

This is an executable shell template, not just documentation. Running it invokes the inference.sh CLI with user-provided URLs and browser sessions.

Factores de riesgo

⚙️ Comandos externos (163)

🌐 Acceso a red (53)

📁 Acceso al sistema de archivos (15)

references/authentication.md:5 references/authentication.md:263 references/commands.md:3 references/commands.md:59 references/proxy-support.md:5 references/proxy-support.md:272 references/snapshot-refs.md:5 references/video-recording.md:5 references/video-recording.md:239 templates/authenticated-session.sh:30 templates/capture-workflow.sh:28 templates/capture-workflow.sh:78 templates/capture-workflow.sh:81 templates/capture-workflow.sh:149 templates/form-automation.sh:23

Patrones detectados

Screen capture upload

Auditado por: codex Ver historial de auditorías →

Compartir y citar este informe

Comparte el informe de evaluación versionado, la insignia neutral, la tarjeta insertable y las citas. Skillstore presenta evidencias sin decidir si este Skill es seguro.

Abrir el informe versionado

Copiar enlace del informe

https://skillstore.io/skills/tul-sh-agent-browser/audits/4?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_report

Insignia Markdown

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

Insignia HTML

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

Tarjeta para insertar

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

Citas académicas (APA · BibTeX · CFF)

Cita APA

tul-sh. (2026). agent-browser security audit report (audit version 4) [Author version unspecified]. Skillstore. https://skillstore.io/skills/tul-sh-agent-browser/audits/4

Cita BibTeX

@techreport{tul-sh-tul-sh-agent-browser-2026,
  author = {tul-sh},
  title = {agent-browser security audit report (audit version 4)},
  institution = {Skillstore},
  year = {2026},
  number = {4},
  url = {https://skillstore.io/skills/tul-sh-agent-browser/audits/4},
  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-browser security audit report (audit version 4)"
version: "unspecified"
type: report
authors:
  - name: "tul-sh"
date-released: "2026-07-07"
url: "https://skillstore.io/skills/tul-sh-agent-browser/audits/4"
identifiers:
  - type: other
    value: "skillstore:tul-sh-agent-browser:audit:4"
    description: "Skillstore immutable audit report identifier"

Comparar variantes

19 variantes instalables

Cada autor sigue siendo un skill instalable independiente. La variante recomendada se ordena según la evidencia de Skillstore.

Por qué esta variante es la primera

Mayor puntuación de Skillstore

Autor / fuente Puntuación de Skillstore Confianza de la evidencia Uso en Skillstore Actualizado

supercent-io Recomendado

supercent-io-agent-browser

Puntuación de Skillstore 78

Confianza de la evidencia Alto

Uso en Skillstore 37

Actualizado

2026-09-09

101-skills

101-skills-agent-browser

Puntuación de Skillstore 74

Confianza de la evidencia Medio

Uso en Skillstore 4

Actualizado

2026-09-09

vercel-labs

vercel-labs-agent-browser

Puntuación de Skillstore 38

Confianza de la evidencia Alto

Uso en Skillstore 72

Actualizado

2026-09-09

qu-skills

qu-skills-agent-browser

Puntuación de Skillstore 38

Confianza de la evidencia Medio

Uso en Skillstore 42

Actualizado

2026-09-09

tul-sh Actual

tul-sh-agent-browser

Puntuación de Skillstore 38

Confianza de la evidencia Medio

Uso en Skillstore 38

Actualizado

2026-09-09

am-will

am-will-agent-browser

Puntuación de Skillstore 38

Confianza de la evidencia Medio

Uso en Skillstore 37

Actualizado

2026-09-09

inference-sh-8

inference-sh-8-agent-browser

Puntuación de Skillstore 38

Confianza de la evidencia Medio

Uso en Skillstore 26

Actualizado

2026-09-09

skillssh

skillssh-agent-browser

Puntuación de Skillstore 38

Confianza de la evidencia Medio

Uso en Skillstore 18

Actualizado

2026-09-09

toolshell

toolshell-agent-browser

Puntuación de Skillstore 38

Confianza de la evidencia Medio

Uso en Skillstore 17

Actualizado

2026-09-09

inference-sh-9

inference-sh-9-agent-browser

Puntuación de Skillstore 38

Confianza de la evidencia Medio

Uso en Skillstore 16

Actualizado

2026-09-09

inferen-sh

inferen-sh-agent-browser

Puntuación de Skillstore 38

Confianza de la evidencia Medio

Uso en Skillstore 11

Actualizado

2026-09-09

inferencesh

inferencesh-agent-browser

Puntuación de Skillstore 38

Confianza de la evidencia Medio

Uso en Skillstore 11

Actualizado

2026-09-09

inference-sh-skills

inference-sh-skills-agent-browser

Puntuación de Skillstore 38

Confianza de la evidencia Medio

Uso en Skillstore 10

Actualizado

2026-09-09

inference-shell

inference-shell-agent-browser

Puntuación de Skillstore 38

Confianza de la evidencia Medio

Uso en Skillstore 10

Actualizado

2026-09-09

inference-sh-0

inference-sh-0-agent-browser

Puntuación de Skillstore 38

Confianza de la evidencia Medio

Uso en Skillstore 10

Actualizado

2026-09-09

infsh-skills

infsh-skills-agent-browser

Puntuación de Skillstore 38

Confianza de la evidencia Medio

Uso en Skillstore 9

Actualizado

2026-09-09

inf-sh

inf-sh-agent-browser

Puntuación de Skillstore 38

Confianza de la evidencia Medio

Uso en Skillstore 8

Actualizado

2026-09-09

inference-skills

inference-skills-agent-browser

Puntuación de Skillstore 38

Confianza de la evidencia Medio

Uso en Skillstore 8

Actualizado

2026-09-09

halt-catch-fire

halt-catch-fire-agent-browser

Puntuación de Skillstore 38

Confianza de la evidencia Medio

Uso en Skillstore 4

Actualizado

2026-09-09

Puntuación de Skillstore

Por qué esta puntuación Confianza de la evidencia: Medio

Arquitectura

Mantenibilidad

Contenido

Comunidad

Cumplimiento de la especificación

Lo que puedes crear

Comprobaciones automatizadas de QA

Ejecutar flujos repetibles del navegador, capturar screenshots y grabar fallos para su revisión.

Captura de investigación

Recopilar texto visible de la página, enlaces, screenshots y metadatos de páginas de sitios web públicos.

Tareas operativas de formularios

Rellenar formularios conocidos, verificar resultados de envío y conservar un registro de sesión para flujos de trabajo internos.

Prueba estos prompts

Abrir e inspeccionar una página

Use agent-browser to open this URL, list the visible interactive elements, and summarize the page state.

Rellenar un formulario público

Open this form, identify the field refs, fill the provided non-sensitive test data, submit it, and verify the result.

Capturar evidencia de un bug

Reproduce this browser issue, take screenshots before and after the key action, and record the session if no secrets are visible.

Extraer contenido estructurado de la página

Open the target page, snapshot current refs, run safe JavaScript to extract headings and links, and save a screenshot for context.

Mejores prácticas

Usar solo cuentas de prueba o credenciales aprobadas para navegación autenticada.
Deshabilitar la grabación cuando las páginas puedan mostrar contraseñas, tokens, datos personales o contenido privado de clientes.
Volver a ejecutar snapshot después de cada navegación antes de usar referencias de elementos.

Evitar

No extraer ni imprimir cookies a menos que el usuario autorice explícitamente esa acción.
No enrutar tráfico a través de proxies para eludir controles de acceso o políticas del sitio.
No subir archivos locales a menos que tanto la ruta del archivo como el destino estén aprobados.

Preguntas frecuentes

¿Qué automatiza esta skill?

Automatiza la navegación del navegador, la interacción con elementos, screenshots, grabación de video, uploads y JavaScript a nivel de página mediante inference.sh.

¿Necesita un servicio externo?

Sí. Depende de la CLI de inference.sh y de la app alojada agent-browser.

¿Puede manejar páginas de inicio de sesión?

Sí, pero los usuarios deben proporcionar credenciales aprobadas y evitar grabar o registrar campos sensibles.

¿Puede hacer scraping de páginas privadas?

Puede extraer contenido visible para la sesión del navegador. Los usuarios deben tener permiso para la página objetivo.

¿Por qué son riesgosos los screenshots y el video?

Pueden capturar credenciales, tokens, datos personales, páginas internas o información de clientes.

¿Debe ejecutarse esta skill en sitios no confiables?

Usar con precaución. Las páginas no confiables pueden influir en el contenido visible, la navegación y los datos extraídos.