📦

vibe-kanban

Name: vibe-kanban
Author: supercent-io

v1.2.0 Revisión del contenido r1 Crítico 🌐 Acceso a red🔑 Variables de entorno📁 Acceso al sistema de archivos⚙️ Comandos externos

Coordina agentes de codificación de IA con Kanban

El trabajo paralelo de codificación con IA es difícil de rastrear entre ramas, logs y pull requests. Esta skill proporciona un flujo de trabajo Kanban para asignar agentes, aislar worktrees, revisar diffs y preparar PRs.

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 "vibe-kanban" from https://skillstore.io/skills/supercent-io-vibe-kanban.md and its manifest at https://skillstore.io/api/skills/supercent-io-vibe-kanban/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/supercent-io-vibe-kanban.md Manifest firmado GET /api/skills/supercent-io-vibe-kanban/manifest Lockfile firmado GET /api/skills/supercent-io-vibe-kanban/lockfile

Pruébalo

Usando "vibe-kanban". Crea un plan de tablero para una actualización del flujo de pagos.

Resultado esperado:

Un conjunto de tarjetas pequeñas para frontend, backend, pruebas de integración y documentación.
Asignaciones de agentes recomendadas para cada tarjeta.
Puntos de control de revisión para logs, diffs y preparación del pull request.

Usando "vibe-kanban". Comprueba si mi servidor Vibe Kanban está listo.

Resultado esperado:

Un breve informe de estado que cubra la accesibilidad del servidor, el número de tarjetas, el estado de worktrees y la disponibilidad de MCP.
Siguientes pasos claros si falta el servidor, el endpoint MCP o la CLI requerida.

Usando "vibe-kanban". Prepara una configuración MCP segura para Codex.

Resultado esperado:

Una revisión de la ubicación objetivo de la configuración de Codex.
El comando esperado del servidor MCP y el puerto local.
Advertencias sobre cambios de configuración de confianza y variables de entorno.

Auditoría de seguridad

Crítico

v4 • 7/7/2026 Abrir el informe versionado

The audit confirmed a critical Docker deployment risk from combining an unpinned container image with a host Docker socket mount. Additional confirmed risks include sourcing .env as shell code, persistent hidden agent configuration writes, secret-printing troubleshooting guidance, unpinned npx execution, and destructive cleanup behavior. Most Markdown backtick, localhost URL, loopback IP, and placeholder secret detections were false positives in documentation context.

Archivos escaneados

1,394

Líneas analizadas

Elementos de revisión

Falsos positivos ignorados

Preocupaciones de seguridad confirmadas (11)

Crítico

templates/docker-compose.yml:17

Docker socket access

- /var/run/docker.sock:/var/run/docker.sock

The Docker Compose template mounts /var/run/docker.sock into the container. A compromised container or dependency could control the host Docker daemon and escalate to host-level access.

Crítico

templates/docker-compose.yml:5 templates/docker-compose.yml:17

Unpinned Docker Image with Host Docker Access

The Docker template uses the latest Vibe Kanban image while also mounting the host Docker socket. A future or compromised image release would receive direct access to the host Docker daemon.

The compose file directly uses an unpinned latest image and mounts /var/run/docker.sock. The combination creates a clear supply-chain path to host-level control.

Alto

scripts/start.sh:78

Environment file access

# .env 파일 로드 (존재하는 경우)

The start script loads a local .env file and line 82 sources it as shell code. A malicious or unreviewed .env can execute commands and expose secrets when the wrapper runs.

Alto

scripts/start.sh:79

Environment file access

if [[ -f ".env" ]]; then

The start script loads a local .env file and line 82 sources it as shell code. A malicious or unreviewed .env can execute commands and expose secrets when the wrapper runs.

Alto

scripts/start.sh:80

Environment file access

info ".env 파일 로드 중..."

The start script loads a local .env file and line 82 sources it as shell code. A malicious or unreviewed .env can execute commands and expose secrets when the wrapper runs.

Alto

scripts/start.sh:82

Environment file access

source .env

The start script loads a local .env file and line 82 sources it as shell code. A malicious or unreviewed .env can execute commands and expose secrets when the wrapper runs.

Alto

SKILL.md:124

Environment file access

Set in `.env` file before starting the server.

The skill tells users to place configuration in a .env file before startup. In this package, the start script sources .env as shell, so the guidance can lead to executing untrusted env-file content.

Alto

SKILL.md:70-75 scripts/start.sh:109-113

Unpinned npx Package Execution

The skill repeatedly starts Vibe Kanban through npx without pinning a package version or checksum, including the executable start wrapper. This creates supply-chain risk because a future package release or dependency compromise would run in the user workspace.

The command examples and start script directly invoke npx vibe-kanban without a pinned version. npx execution is legitimate for this tool but high impact because it runs external package code locally.

Mostrar los 11 problemas confirmados

Alto

scripts/start.sh:78-82

Environment File Sourced as Shell Code

The start script loads .env with shell source instead of parsing key-value pairs. A crafted .env can execute arbitrary shell commands before Vibe Kanban starts.

Line 82 directly uses source .env, which executes shell syntax from the file. This is a clear code execution risk when .env content is untrusted or copied from another source.

Alto

SKILL.md:417-423

Auto-Approved Agent Actions Increase Blast Radius

The skill documents Auto Approve as enabled by default for agents. In a board that launches coding agents in git worktrees, default approval can allow broad filesystem and command actions before human review.

The documentation explicitly states Auto Approve is on by default and the surrounding workflow runs agent CLIs against repositories. The exact runtime permissions depend on agent configuration, so confidence is high but not absolute.

Medio

scripts/cleanup.sh:70-88

Cleanup Script Can Force Delete Worktrees and Branches

The cleanup helper identifies Vibe Kanban worktrees and can force-remove worktrees and delete branches. A parsing error or broad --all run can remove local work that has not been reviewed.

The script loops over detected worktrees and runs forced removal and branch deletion commands. It supports dry-run, but the destructive path is present and automated.

Elementos de revisión de capacidades (21)

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

scripts/mcp-setup.sh:61

Hidden file access

CLAUDE_CONFIG="$HOME/.claude/claude_desktop_config.json"

The MCP setup script targets hidden Claude configuration under the user home directory. Writing agent MCP configuration is persistent and can change future tool execution behavior.

Alto

scripts/mcp-setup.sh:62

Hidden file access

CLAUDE_DIR="$HOME/.claude"

The MCP setup script targets hidden Claude configuration under the user home directory. Writing agent MCP configuration is persistent and can change future tool execution behavior.

Alto

scripts/mcp-setup.sh:97

Hidden file access

CODEX_CONFIG="$HOME/.codex/config.toml"

The MCP setup script targets hidden Codex configuration under the user home directory. Writing agent MCP configuration is persistent and can change future tool execution behavior.

Alto

scripts/mcp-setup.sh:98

Hidden file access

CODEX_DIR="$HOME/.codex"

The MCP setup script targets hidden Codex configuration under the user home directory. Writing agent MCP configuration is persistent and can change future tool execution behavior.

Alto

SKILL.md:140

Hidden file in home directory

`~/.claude/settings.json` or project `.mcp.json`: