AgentDrive

Cross-agent artifact workspace with provenance across Claude Code, Codex, Cursor, LangGraph.

¿Debería usar esto?

Calidad y seguridad

A
Calidad de la descripción
95%
Integridad del esquema
96%
Calidad de los nombres
80%
Riesgo de envenenamiento
100%
Coincidencia de permisos
100%
Cumplimiento del protocolo
100%

Hallazgos (3)

  • LOWTool 'drive_org' description lacks action verben drive_org
  • LOWTool 'drive_versions' description lacks action verben drive_versions
  • LOWTool 'drive_sandbox_commit' description lacks action verben drive_sandbox_commit

Basado en el análisis automatizado de las definiciones de herramientas y el cumplimiento del protocolo.

Costo de contexto

~2,506Tokens (definiciones de herramientas)
~1.1 KBTamaño de respuesta típico
Impacto moderado en la atención (1.96% del contexto de 128k)

Este es el número aproximado de tokens que se consumen cada vez que las herramientas del servidor se cargan en el contexto de un modelo. Los recuentos más altos reducen la atención disponible para otras tareas.

Instalar

Instalación con un clic

Agrega esto a tu archivo `claude_desktop_config.json`:

{
  "mcpServers": {
    "agentdrive": {
      "url": "https://mcp.agentdrive.sh/mcp"
    }
  }
}

Puntos de conexión remotos

https://mcp.agentdrive.sh/mcpstreamable-http

Qué puede hacer

Inventario de herramientas

Herramientas (20)

🟢 Solo lectura🟡 Escritura🔴 Eliminación⚪ Desconocido
🟡drive_write(workspace_id, path, content, message)

Write a markdown file to a workspace. Creates or updates the file with automatic versioning.

Esquema de entrada

{
  "type": "object",
  "properties": {
    "workspace_id": {
      "type": "string",
      "description": "Workspace ID to write to"
    },
    "path": {
      "type": "string",
      "description": "File path (e.g. \"notes/meeting.md\")"
    },
    "content": {
      "type": "string",
      "description": "File content (markdown)"
    },
    "message": {
      "type": "string",
      "description": "Optional version message"
    }
  },
  "required": [
    "workspace_id",
    "path",
    "content"
  ],
  "additionalProperties": false
}
🟢drive_read(workspace_id, file_id)

Read a file's content from a workspace.

Esquema de entrada

{
  "type": "object",
  "properties": {
    "workspace_id": {
      "type": "string",
      "description": "Workspace ID"
    },
    "file_id": {
      "type": "string",
      "description": "File ID to read"
    }
  },
  "required": [
    "workspace_id",
    "file_id"
  ],
  "additionalProperties": false
}
🟢drive_list(workspace_id, prefix)

List files in a workspace, optionally filtered by path prefix.

Esquema de entrada

{
  "type": "object",
  "properties": {
    "workspace_id": {
      "type": "string",
      "description": "Workspace ID"
    },
    "prefix": {
      "type": "string",
      "description": "Path prefix filter (e.g. \"notes/\")"
    }
  },
  "required": [
    "workspace_id"
  ],
  "additionalProperties": false
}
🟡drive_share(workspace_id, permission, file_id, expires_in, max_uses, ...)

Generate a share link for a workspace or specific file. Optionally email-gate the link so only a specific person can access it. Humans open this link to view, comment, or edit files.

Esquema de entrada

{
  "type": "object",
  "properties": {
    "workspace_id": {
      "type": "string",
      "description": "Workspace ID"
    },
    "permission": {
      "type": "string",
      "enum": [
        "view",
        "comment",
        "edit"
      ],
      "description": "Permission level for the share link"
    },
    "file_id": {
      "type": "string",
      "description": "Scope to a specific file (optional)"
    },
    "expires_in": {
      "type": "number",
      "description": "Expiry in seconds (optional)"
    },
    "max_uses": {
      "type": "number",
      "description": "Maximum number of uses (optional)"
    },
    "email": {
      "type": "string",
      "description": "Email-gate: only this email can access the link (requires email verification). Leave empty for public access."
    }
  },
  "required": [
    "workspace_id",
    "permission"
  ],
  "additionalProperties": false
}
🟡drive_comment(workspace_id, file_id, body, line_start, line_end)

Add a comment to a file.

Esquema de entrada

{
  "type": "object",
  "properties": {
    "workspace_id": {
      "type": "string",
      "description": "Workspace ID"
    },
    "file_id": {
      "type": "string",
      "description": "File ID to comment on"
    },
    "body": {
      "type": "string",
      "description": "Comment text"
    },
    "line_start": {
      "type": "number",
      "description": "Start line number (optional)"
    },
    "line_end": {
      "type": "number",
      "description": "End line number (optional)"
    }
  },
  "required": [
    "workspace_id",
    "file_id",
    "body"
  ],
  "additionalProperties": false
}
🟢drive_feedback(workspace_id, file_id)

Read human comments on a file. Use this to check if humans have left feedback.

Esquema de entrada

{
  "type": "object",
  "properties": {
    "workspace_id": {
      "type": "string",
      "description": "Workspace ID"
    },
    "file_id": {
      "type": "string",
      "description": "File ID to read comments from"
    }
  },
  "required": [
    "workspace_id",
    "file_id"
  ],
  "additionalProperties": false
}
🟢drive_activity(workspace_id)

Get recent activity in a workspace (file writes, comments, shares).

Esquema de entrada

{
  "type": "object",
  "properties": {
    "workspace_id": {
      "type": "string",
      "description": "Workspace ID"
    }
  },
  "required": [
    "workspace_id"
  ],
  "additionalProperties": false
}
⚪drive_invite(email, role)

Invite a human to your organization by email. They will receive a verification code to join.

Esquema de entrada

{
  "type": "object",
  "properties": {
    "email": {
      "type": "string",
      "description": "Email address of the human to invite to your organization"
    },
    "role": {
      "type": "string",
      "enum": [
        "owner",
        "admin",
        "member"
      ],
      "default": "member",
      "description": "Role in the organization (default: member)"
    }
  },
  "required": [
    "email"
  ],
  "additionalProperties": false
}
🔴drive_roles(workspace_id, action, email, permission)

Manage per-workspace role delegation. Assign, list, or remove access for specific emails (view/comment/edit).

Esquema de entrada

{
  "type": "object",
  "properties": {
    "workspace_id": {
      "type": "string",
      "description": "Workspace ID"
    },
    "action": {
      "type": "string",
      "enum": [
        "assign",
        "list",
        "remove"
      ],
      "description": "Action: assign a role, list all roles, or remove a role"
    },
    "email": {
      "type": "string",
      "description": "Email address (required for assign/remove)"
    },
    "permission": {
      "type": "string",
      "enum": [
        "view",
        "comment",
        "edit"
      ],
      "description": "Permission level (required for assign)"
    }
  },
  "required": [
    "workspace_id",
    "action"
  ],
  "additionalProperties": false
}
🟢drive_org

View your organization details: agents, members, plan, and usage.

Esquema de entrada

{
  "type": "object",
  "properties": {},
  "additionalProperties": false
}
🔴drive_delete(workspace_id, file_id)

Delete a file from a workspace.

Esquema de entrada

{
  "type": "object",
  "properties": {
    "workspace_id": {
      "type": "string",
      "description": "Workspace ID"
    },
    "file_id": {
      "type": "string",
      "description": "File ID to delete"
    }
  },
  "required": [
    "workspace_id",
    "file_id"
  ],
  "additionalProperties": false
}
🟢drive_versions(workspace_id, file_id)

View version history for a file in a workspace.

Esquema de entrada

{
  "type": "object",
  "properties": {
    "workspace_id": {
      "type": "string",
      "description": "Workspace ID"
    },
    "file_id": {
      "type": "string",
      "description": "File ID"
    }
  },
  "required": [
    "workspace_id",
    "file_id"
  ],
  "additionalProperties": false
}
🟢drive_search(query, workspace_id, top_k, filter)

Semantic search across your org's artifacts in AgentDrive. Returns the top-N most relevant files. Use this BEFORE writing similar artifacts to avoid duplication and surface prior context.

Esquema de entrada

{
  "type": "object",
  "properties": {
    "query": {
      "type": "string",
      "minLength": 1,
      "maxLength": 2000,
      "description": "Search query"
    },
    "workspace_id": {
      "type": "string",
      "description": "Scope to a single workspace; omit for org-wide search"
    },
    "top_k": {
      "type": "integer",
      "exclusiveMinimum": 0,
      "maximum": 50,
      "description": "Max hits to return (default 10, max 50)"
    },
    "filter": {
      "type": "object",
      "properties": {
        "agent_platform": {
          "type": "string",
          "enum": [
            "claude-code",
            "codex",
            "cursor",
            "langgraph",
            "crewai",
            "autogen",
            "hermes",
            "openclaw",
            "unknown"
          ]
        },
        "agent_id": {
          "type": "string"
        },
        "path_prefix": {
          "type": "string"
        }
      },
      "additionalProperties": false
    }
  },
  "required": [
    "query"
  ],
  "additionalProperties": false
}
🟢drive_recall(query, top_k)

Recall persistent facts and preferences your agents have written. In v1 this returns semantic search results across artifacts; v1.5 will narrow to auto-extracted facts (preferences, decisions, key entities).

Esquema de entrada

{
  "type": "object",
  "properties": {
    "query": {
      "type": "string",
      "minLength": 1,
      "maxLength": 2000,
      "description": "Recall query — what fact, preference, or decision are you trying to remember?"
    },
    "top_k": {
      "type": "integer",
      "exclusiveMinimum": 0,
      "maximum": 20,
      "description": "Max results (default 5, max 20)"
    }
  },
  "required": [
    "query"
  ],
  "additionalProperties": false
}
⚪drive_log_run(task, status, run_id, artifacts_referenced, artifacts_produced, ...)

Log an agent run to AgentDrive's provenance store. Captures task, status, referenced/produced artifacts, and cross-platform identity (which agent on which tool). Used for retrospective context retrieval and team-wide audit.

Esquema de entrada

{
  "type": "object",
  "properties": {
    "task": {
      "type": "string",
      "minLength": 1,
      "description": "What the agent did or is doing (free text)"
    },
    "status": {
      "type": "string",
      "enum": [
        "success",
        "failure",
        "partial",
        "in_progress"
      ],
      "description": "Run outcome (default success)"
    },
    "run_id": {
      "type": "string",
      "description": "Existing run ID to update; omit to let server allocate"
    },
    "artifacts_referenced": {
      "type": "array",
      "items": {
        "type": "string"
      },
      "description": "Paths of artifacts the run read"
    },
    "artifacts_produced": {
      "type": "array",
      "items": {
        "type": "string"
      },
      "description": "Paths of artifacts the run wrote"
    },
    "duration_ms": {
      "type": "integer",
      "minimum": 0,
      "description": "Wall-clock duration in milliseconds"
    },
    "metadata": {
      "type": "object",
      "additionalProperties": {},
      "description": "Free-form key/value metadata"
    }
  },
  "required": [
    "task"
  ],
  "additionalProperties": false
}
⚪drive_context(task, workspace_id, limit)

Auto-pull the most relevant artifacts from prior agent sessions for the task you are about to do. Combines semantic relevance with recency. Call this at the START of any non-trivial task to ground the agent in prior context — replaces manual copy-paste of artifacts between sessions.

Esquema de entrada

{
  "type": "object",
  "properties": {
    "task": {
      "type": "string",
      "minLength": 1,
      "maxLength": 2000,
      "description": "What you are about to do — used to rank prior artifacts by relevance"
    },
    "workspace_id": {
      "type": "string",
      "description": "Scope to a single workspace; omit for org-wide context"
    },
    "limit": {
      "type": "integer",
      "exclusiveMinimum": 0,
      "maximum": 20,
      "description": "Max artifacts to return (default 5, max 20)"
    }
  },
  "required": [
    "task"
  ],
  "additionalProperties": false
}
⚪drive_sandbox_clone(workspace_id, url, branch, depth)

[EXPERIMENTAL] Clone a git repo into a per-workspace Linux sandbox. Requires Cloudflare Containers on the AgentDrive worker; gracefully degrades to a 503 with "sandbox not configured" otherwise.

Esquema de entrada

{
  "type": "object",
  "properties": {
    "workspace_id": {
      "type": "string",
      "description": "Workspace to clone into. The sandbox persists across MCP sessions per workspace."
    },
    "url": {
      "type": "string",
      "format": "uri",
      "description": "Git repository URL to clone"
    },
    "branch": {
      "type": "string",
      "description": "Branch to check out (default: repo default)"
    },
    "depth": {
      "type": "integer",
      "exclusiveMinimum": 0,
      "maximum": 50,
      "description": "Shallow clone depth (default: full clone)"
    }
  },
  "required": [
    "workspace_id",
    "url"
  ],
  "additionalProperties": false
}
⚪drive_sandbox_exec(workspace_id, cmd, cwd, timeout_ms)

[EXPERIMENTAL] Run a shell command inside the per-workspace Linux sandbox. Captures stdout/stderr (64 KB cap), exit code, duration. Requires Cloudflare Containers; degrades to 503 otherwise.

Esquema de entrada

{
  "type": "object",
  "properties": {
    "workspace_id": {
      "type": "string",
      "description": "Workspace whose sandbox to run in"
    },
    "cmd": {
      "type": "string",
      "minLength": 1,
      "maxLength": 2048,
      "description": "Shell command to run inside the per-workspace Linux sandbox"
    },
    "cwd": {
      "type": "string",
      "description": "Working directory (default: workspace root)"
    },
    "timeout_ms": {
      "type": "integer",
      "exclusiveMinimum": 0,
      "maximum": 60000,
      "description": "Max runtime in ms (default 30000, hard cap 60000)"
    }
  },
  "required": [
    "workspace_id",
    "cmd"
  ],
  "additionalProperties": false
}
🟢drive_sandbox_read(workspace_id, path)

[EXPERIMENTAL] Read a file from the per-workspace sandbox filesystem (build artifacts, test output, logs). Requires Cloudflare Containers; degrades to 503 otherwise.

Esquema de entrada

{
  "type": "object",
  "properties": {
    "workspace_id": {
      "type": "string",
      "description": "Workspace whose sandbox to read from"
    },
    "path": {
      "type": "string",
      "description": "Path inside the sandbox (e.g., src/index.ts or /tmp/build.log)"
    }
  },
  "required": [
    "workspace_id",
    "path"
  ],
  "additionalProperties": false
}
⚪drive_sandbox_commit(workspace_id, message, author_name, author_email)

[EXPERIMENTAL] Stage + commit the current sandbox worktree. Surfaces as an AgentDrive artifact at commits/<sha>.json. Requires Cloudflare Containers; degrades to 503 otherwise.

Esquema de entrada

{
  "type": "object",
  "properties": {
    "workspace_id": {
      "type": "string",
      "description": "Workspace whose sandbox to commit"
    },
    "message": {
      "type": "string",
      "minLength": 1,
      "maxLength": 500,
      "description": "Commit message"
    },
    "author_name": {
      "type": "string",
      "description": "Override git author name (default: agent identity)"
    },
    "author_email": {
      "type": "string",
      "format": "email",
      "description": "Override git author email (default: agent identity)"
    }
  },
  "required": [
    "workspace_id",
    "message"
  ],
  "additionalProperties": false
}

Prompts recomendados

search_research
Search for information about [topic] using AgentDrive
Herramientas esperadas: drive_search
find_specific
Find [specific item] using AgentDrive
Herramientas esperadas: drive_search
retrieve_data
Get details about [item] from AgentDrive
Herramientas esperadas: drive_read
fetch_info
Fetch [information type] using AgentDrive
Herramientas esperadas: drive_read
list_items
List all [items] available in AgentDrive
Herramientas esperadas: drive_list

Comunidad

Califica este servidor

Evidencia

Observaciones recientes

verificadoversión no registrada20 herramientas
verificadoversión no registrada20 herramientas