📦

spawn

Name: spawn
Author: Consiliency

Revisión del contenido r1 Crítico 🔑 Variables de entorno⚙️ Comandos externos📁 Acceso al sistema de archivos

Generar agentes y comandos de terminal

Los desarrolladores a menudo necesitan agentes interactivos o tareas de CLI de larga duración sin bloquear su sesión principal. Esta skill abre sesiones de terminal para agentes de programación y comandos genéricos, mientras documenta la captura de salida.

Compatible con: Claude Codex Code(CC)

⚠️ 38 Deficiente

No instales este Skill automáticamente.

La política canónica exige la revisión de un operador antes de cualquier acción de instalación.

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 "spawn" from https://skillstore.io/skills/consiliency-spawn.md and its manifest at https://skillstore.io/api/skills/consiliency-spawn/manifest. Verify the artifact. Do not auto-install. Inspect the skill and report your findings, then wait for an operator or manual installation decision.

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/consiliency-spawn.md Manifest firmado GET /api/skills/consiliency-spawn/manifest Lockfile firmado GET /api/skills/consiliency-spawn/lockfile

Pruébalo

Usando "spawn". Genera una terminal para una conversión de video larga.

Resultado esperado:

Se abre una nueva terminal para el comando. El asistente informa si la salida de depuración se captura y dónde encontrarla.

Usando "spawn". Inicia Codex para revisar el conjunto de pruebas fallidas.

Resultado esperado:

La skill selecciona el cookbook de Codex, crea el comando de CLI, abre una terminal y devuelve la ruta de salida capturada.

Usando "spawn". Abre un flujo de inicio de sesión para Cursor Agent.

Resultado esperado:

Se abre una terminal para el comando de autenticación de Cursor. El llamador espera a que el flujo de inicio de sesión finalice o agote el tiempo de espera.

Auditoría de seguridad

Crítico

v11 • 19/7/2026 Abrir el informe versionado

Most static alerts in Markdown, comments, UUID generation, and constant /proc reads are false positives. Confirmed risks include unsanitized shell command execution and shared temporary files that can expose agent output or generated scripts. Cookbook defaults also disable approval and sandbox controls, which can grant unrestricted system access.

Archivos escaneados

1,587

Líneas analizadas

Elementos de revisión

Falsos positivos ignorados

Preocupaciones de seguridad confirmadas (2)

Crítico

agent/cookbook/codex-cli.md:45-46 agent/cookbook/cursor-cli.md:40-43

Agent Safety Controls Disabled by Default

The cookbooks direct agents to bypass sandbox and approval controls or force command approval. Spawned agents can modify files and execute commands without confirmation.

The Codex cookbook explicitly says to always bypass approvals and the sandbox, while the Cursor cookbook directs force approval. Both explain that normal confirmations are removed.

Medio

agent/SKILL.md:57-62 agent/prompts/fork-summary-user-prompt.md:1-2

Conversation History Shared With External Providers

The summary workflow copies conversation history into a prompt for a separate provider agent without a required secret-redaction or disclosure review.

The workflow explicitly replaces summary fields with conversation history and passes them to another agent. The user requests a summary, but no redaction step protects embedded secrets.

Elementos de revisión de capacidades (28)

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

Crítico

agent/fork_terminal.py:176

Python subprocess.run

subprocess.run(f"cmd.exe /c 'start cmd /k wsl -- {command}'", shell=True)

The function interpolates the caller-controlled command into a cmd.exe string and executes it with shell=True. Shell metacharacters can run arbitrary commands.

Crítico

agent/fork_terminal.py:216

Python subprocess.Popen

subprocess.Popen(wrapped_cmd, shell=True, cwd=cwd)

A caller-controlled command is placed in wrapped_cmd and executed with shell=True. This is direct arbitrary shell execution without validation or argument separation.

Crítico

agent/fork_terminal.py:222

Python subprocess.Popen

subprocess.Popen(wrapped_cmd, shell=True, cwd=cwd)

A caller-controlled command is placed in wrapped_cmd and executed with shell=True while capturing debug output. Shell metacharacters can execute additional commands.

Crítico

agent/fork_terminal.py:227

Python subprocess.Popen

subprocess.Popen(command, shell=True, cwd=cwd)

The command parameter is passed directly to subprocess.Popen with shell=True. Any untrusted command text receives arbitrary shell execution.

Crítico

agent/fork_terminal.py:176

Windows cmd.exe

subprocess.run(f"cmd.exe /c 'start cmd /k wsl -- {command}'", shell=True)

cmd.exe receives a formatted string containing the caller-controlled command through shell=True. Nested shell parsing creates a direct command injection path.

Alto

agent/fork_terminal.py:58

Python subprocess.run

subprocess.run(

The osascript process receives AppleScript containing an interpolated shell command and working directory. Unescaped command text can execute arbitrary shell or AppleScript content.

Alto

agent/fork_terminal.py:93

Python subprocess.Popen

subprocess.Popen(wt_cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE)

Although Popen receives an argument list, powershell.exe interprets ps_cmd, which embeds the caller-controlled command. The list form does not prevent injection inside the PowerShell command string.

Alto

agent/fork_terminal.py:161

Python subprocess.Popen

subprocess.Popen(wt_cmd)

This launches a generated WSL script whose body contains the caller-controlled command without validation. The argument list protects process selection but not the script contents.

Alto

agent/fork_terminal.py:196

Python subprocess.Popen

subprocess.Popen(tmux_cmd)

The tmux command invokes bash with a constructed command string that contains caller-controlled input. Shell operators in that input execute with the skill process permissions.

Alto

agent/fork_terminal.py:90

PowerShell invocation

"wt.exe", "-w", "new", "powershell.exe",

PowerShell is launched with a command string assembled from caller-controlled command text. PowerShell parses that text and can execute injected commands.

Alto

agent/fork_terminal.py:140

Unix shell invocation

script_content = f'''#!/bin/bash -li

The generated Bash script embeds the caller-controlled command before executing it. No validation or argument-level quoting limits shell operators.

Alto

agent/fork_terminal.py:146

Unix shell invocation

script_content = f'''#!/bin/bash -li

The debug-capture Bash script embeds the caller-controlled command directly. Arbitrary shell syntax is executed when the script starts.

Alto

agent/fork_terminal.py:152

Unix shell invocation

script_content = f'''#!/bin/bash -li

The no-capture Bash script places the command parameter into executable script content without validation. This intentionally permits arbitrary shell execution.

Alto

agent/fork_terminal.py:137

Temp directory access

script_file = f"/tmp/fork-script-{output_id}.sh"

An executable script containing the requested command is created at a shared /tmp path. The code does not use exclusive creation, private permissions, or cleanup.

Alto

agent/fork_terminal.py:157

Python file write/append

with open(script_file, 'w') as f:

The code writes caller-controlled command text to a shared temporary script without exclusive creation. A local race can replace the target, and the script can disclose prompt content.

Alto

agent/fork_terminal.py:159

Python os file operations

os.chmod(script_file, 0o755)

The generated shared temporary script is made world-readable and executable with mode 0755. It can expose command or prompt contents and remains after execution.

Medio

agent/fork_terminal.py:41

Temp directory access

agent_output_file = f"/tmp/fork-agent-{output_id}.json" if log_agent_output else None

Agent responses are written under shared /tmp using a short identifier and normal shell redirection. The code does not enforce private permissions or remove the file.

Medio

agent/fork_terminal.py:42

Temp directory access

debug_output_file = f"/tmp/fork-debug-{output_id}.txt" if log_to_file else None

Full terminal output can contain prompts, errors, or credentials and is stored under shared /tmp. No private file mode or cleanup is applied.

Medio

agent/fork_terminal.py:43

Temp directory access

done_file = f"/tmp/fork-output-{output_id}.done"

The completion marker uses a shared /tmp path and is later trusted by existence checks. Pre-creation can cause premature completion detection, although the random identifier reduces feasibility.

Medio

agent/fork_terminal.py:134

Temp directory access

agent_output_file = f"/tmp/fork-agent-{output_id}.json" if log_agent_output else None

WSL agent responses are stored in shared /tmp without explicit private permissions or cleanup. Captured responses can contain sensitive repository or conversation data.

Medio

agent/fork_terminal.py:135

Temp directory access

debug_output_file = f"/tmp/fork-debug-{output_id}.txt" if log_to_file else None

WSL debug output is stored in shared /tmp without explicit private permissions or cleanup. Full terminal logs can contain sensitive data.

Medio

agent/fork_terminal.py:136

Temp directory access

done_file = f"/tmp/fork-output-{output_id}.done"

The WSL completion marker is created in shared /tmp and trusted by existence checks. Pre-creation can affect capture integrity, although the random identifier reduces feasibility.

Medio

agent/fork_terminal.py:183

Temp directory access

agent_output_file = f"/tmp/fork-agent-{output_id}.json" if log_agent_output else None

tmux agent output is stored in shared /tmp without explicit private permissions or cleanup. The response can expose repository or conversation content.

Medio

agent/fork_terminal.py:184

Temp directory access

debug_output_file = f"/tmp/fork-debug-{output_id}.txt" if log_to_file else None

tmux debug output is stored in shared /tmp without explicit private permissions or cleanup. Full terminal logs can contain sensitive data.

Medio

agent/fork_terminal.py:185

Temp directory access

done_file = f"/tmp/fork-output-{output_id}.done"

The tmux completion marker uses shared /tmp and is trusted by existence checks. Pre-creation can affect output synchronization, although the random identifier reduces feasibility.

Medio

agent/fork_terminal.py:210

Temp directory access

agent_output_file = f"/tmp/fork-agent-{output_id}.json" if log_agent_output else None

Fallback agent output is stored in shared /tmp without explicit private permissions or cleanup. Captured agent content can be sensitive.

Medio

agent/fork_terminal.py:211

Temp directory access

debug_output_file = f"/tmp/fork-debug-{output_id}.txt" if log_to_file else None

Fallback debug output is stored in shared /tmp without explicit private permissions or cleanup. Terminal output can include sensitive data.

Medio

agent/fork_terminal.py:212

Temp directory access

done_file = f"/tmp/fork-output-{output_id}.done"

The fallback completion marker uses shared /tmp and is later trusted by existence checks. Pre-creation can affect capture integrity, although the random identifier reduces feasibility.

Factores de riesgo

🔑 Variables de entorno (3)

agent/cookbook/cursor-cli.md:128 agent/cookbook/cursor-cli.md:148 agent/fork_terminal.py:96

⚙️ Comandos externos (49)

📁 Acceso al sistema de archivos (23)

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/consiliency-spawn/audits/11?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_report

Insignia Markdown

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

Insignia HTML

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

Tarjeta para insertar

<iframe src="https://skillstore.io/embed/skills/consiliency-spawn.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

Consiliency. (2026). spawn security audit report (audit version 11) [Author version unspecified]. Skillstore. https://skillstore.io/skills/consiliency-spawn/audits/11

Cita BibTeX

@techreport{consiliency-consiliency-spawn-2026,
  author = {Consiliency},
  title = {spawn security audit report (audit version 11)},
  institution = {Skillstore},
  year = {2026},
  number = {11},
  url = {https://skillstore.io/skills/consiliency-spawn/audits/11},
  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: "spawn security audit report (audit version 11)"
version: "unspecified"
type: report
authors:
  - name: "Consiliency"
date-released: "2026-07-19"
url: "https://skillstore.io/skills/consiliency-spawn/audits/11"
identifiers:
  - type: other
    value: "skillstore:consiliency-spawn:audit:11"
    description: "Skillstore immutable audit report identifier"

Puntuación de Skillstore

Por qué esta puntuación Confianza de la evidencia: Medio

Arquitectura

Mantenibilidad

Contenido

Comunidad

Cumplimiento de la especificación

Lo que puedes crear

Ejecutar trabajos de CLI largos

Abrir una terminal separada para comandos de compilación, medios o paquetes mientras el asistente principal sigue trabajando.

Iniciar agentes interactivos de programación

Iniciar una CLI de programación con IA seleccionada con un prompt, elección de modelo y captura de salida opcional.

Gestionar flujos de inicio de sesión de proveedores

Abrir una terminal para la autenticación del proveedor cuando una CLI de programación necesita inicio de sesión interactivo o en navegador.

Prueba estos prompts

Abrir una terminal de compilación

Usa spawn-terminal para abrir una nueva terminal y ejecutar npm run build. Revisa primero la ayuda del comando o el script del paquete.

Iniciar agente Claude Code

Usa spawn-agent con Claude Code para revisar el informe de error actual. Mantén permisos conservadores y devuelve la ruta de salida.

Ejecutar Codex con resumen

Usa spawn-agent con Codex. Incluye un resumen conciso de nuestra tarea actual y luego pídele que inspeccione las pruebas fallidas.

Comparar resultados de agentes

Genera agentes separados para Codex y Cursor para investigar la regresión de rendimiento. Captura salidas limpias y resume los desacuerdos.

Mejores prácticas

Revisa cada comando antes de generarlo en una terminal.
Prefiere modos de agente aislados o basados en confirmación para prompts no confiables.
Elimina los archivos de salida capturados cuando puedan contener datos sensibles.

Evitar

No ejecutes prompts no confiables con omisión de aprobación o acceso completo al sistema.
No pases secretos en prompts, argumentos de comandos ni archivos de salida capturados.
No uses esta skill para comandos simples que puedan ejecutarse de forma segura en la sesión actual.