# workbench_run_python Run Python in an isolated sandbox to process LARGE or paginated tool results without pulling every row into the conversation. Inside the code, call your connected integration tools with call_tool('ext_', {..}). RETURN SHAPE: call_tool ALWAYS returns a dict with a boolean r['success']. On SUCCESS the API's JSON is under r['body'], e.g. {'success': True, 'status': 200, 'body': {'results': [{'title': ...}, ...]}} — so read r['body']['results']. On FAILURE r['success'] is False and r['error'] explains. If unsure of the shape, print(r) once and inspect before extracting. Aggregate/filter/paginate in the sandbox, then assign ONLY the small summary you want back to a variable named result. FIRST discover exact tool slugs with integrations_search_tools, THEN write code that calls them. pandas/numpy available. Agent View of the PolicyLayer registry record for `workbench_run_python`. HTML page: https://policylayer.com/tools/io-github-saloprj-dialogbrain/workbench-run-python ## Facts - Tool: `workbench_run_python` - Server: Dialogbrain (`https://api.dialogbrain.com/mcp`) — https://policylayer.com/tools/io-github-saloprj-dialogbrain.md - Homepage: https://github.com/saloprj/dialogbrain-mcp - Risk category: Execute (High risk) - Registry record: grade F, identity unverified - Server auth posture: open - Server rate-limited: no - Parameters: 2 (1 required) - Recommended policy verdict: Rate-limited ## Parameters | Parameter | Type | Required | Description | | --- | --- | --- | --- | | `code` | string | yes | Python source to execute. call_tool('ext_', {..}) returns the integration's raw dict — HTTP success payload is under r['body'] (e.g. r['body']['result | | `agent_id` | integer | no | Which agent's tool policy the sandbox runs under — this scopes which ext* integrations call_tool may reach (enabled + denied_tools for that agent). Only needed | Parameters from the server's own tool schema. ## Example call (MCP tools/call, JSON-RPC 2.0) ```json { "jsonrpc": "2.0", "id": 1, "method": "tools/call", "params": { "name": "workbench_run_python", "arguments": { "code": "" } } } ``` ## Why workbench_run_python is rated High Executing arbitrary Python code in a sandbox, even isolated, is an Execute-category risk. While the sandbox provides some containment, the tool permits calling other integration tools (WhatsApp, Telegram, Email, voice) from within executed code, potentially allowing an AI agent to chain operations across multiple communication channels and data sources. From the tool's own definition: "Tool explicitly allows to 'Run Python in an isolated sandbox' and states 'Inside the code, call your connected integration tools with `call_tool(...)`'." Risk signals: Accepts freeform code/query input (code) · Bulk/mass operation — affects multiple targets ## Use case AI agents invoke workbench_run_python to trigger actions in Dialogbrain. What it does depends on the arguments the agent supplies, and its effects often reach beyond the immediate call: builds kicked off, notifications sent, workflows started. ## Recommended policy (PolicyLayer) Verdict: **Rate-limited**. Enforced by the PolicyLayer MCP gateway (https://policylayer.com/mcp-gateway) before a call reaches Dialogbrain: ```json { "version": "1", "default": "deny", "tools": { "workbench_run_python": { "limits": [ { "counter": "workbench_run_python_rate", "window": "minute", "max": 10, "scope": "grant" } ] } } } ``` ## Other tools on Dialogbrain (224) - `agents_delete` — Destructive — https://policylayer.com/tools/io-github-saloprj-dialogbrain/agents-delete.md - `agents_trigger_delete` — Destructive — https://policylayer.com/tools/io-github-saloprj-dialogbrain/agents-trigger-delete.md - `ai_filters_delete` — Destructive — https://policylayer.com/tools/io-github-saloprj-dialogbrain/ai-filters-delete.md - `ai_tags_delete` — Destructive — https://policylayer.com/tools/io-github-saloprj-dialogbrain/ai-tags-delete.md - `calendar_delete_event` — Destructive — https://policylayer.com/tools/io-github-saloprj-dialogbrain/calendar-delete-event.md - `collections_delete` — Destructive — https://policylayer.com/tools/io-github-saloprj-dialogbrain/collections-delete.md - `files_delete` — Destructive — https://policylayer.com/tools/io-github-saloprj-dialogbrain/files-delete.md - `folders_delete` — Destructive — https://policylayer.com/tools/io-github-saloprj-dialogbrain/folders-delete.md - `integrations_remove_endpoints` — Destructive — https://policylayer.com/tools/io-github-saloprj-dialogbrain/integrations-remove-endpoints.md - `messages_delete` — Destructive — https://policylayer.com/tools/io-github-saloprj-dialogbrain/messages-delete.md - `notes_delete` — Destructive — https://policylayer.com/tools/io-github-saloprj-dialogbrain/notes-delete.md - `reminder_cancel` — Destructive — https://policylayer.com/tools/io-github-saloprj-dialogbrain/reminder-cancel.md - `tasks_delete` — Destructive — https://policylayer.com/tools/io-github-saloprj-dialogbrain/tasks-delete.md - `threads_delete` — Destructive — https://policylayer.com/tools/io-github-saloprj-dialogbrain/threads-delete.md - `widgets_delete` — Destructive — https://policylayer.com/tools/io-github-saloprj-dialogbrain/widgets-delete.md - `youtube_delete_comment` — Destructive — https://policylayer.com/tools/io-github-saloprj-dialogbrain/youtube-delete-comment.md - `youtube_delete_video` — Destructive — https://policylayer.com/tools/io-github-saloprj-dialogbrain/youtube-delete-video.md - `agent_handoff` — Execute — https://policylayer.com/tools/io-github-saloprj-dialogbrain/agent-handoff.md - `agents_simulate_inbound` — Execute — https://policylayer.com/tools/io-github-saloprj-dialogbrain/agents-simulate-inbound.md - `android_launch_app` — Execute — https://policylayer.com/tools/io-github-saloprj-dialogbrain/android-launch-app.md - `android_shell` — Execute — https://policylayer.com/tools/io-github-saloprj-dialogbrain/android-shell.md - `android_ui_dump` — Execute — https://policylayer.com/tools/io-github-saloprj-dialogbrain/android-ui-dump.md - `background_run` — Execute — https://policylayer.com/tools/io-github-saloprj-dialogbrain/background-run.md - `browser_attach_meet` — Execute — https://policylayer.com/tools/io-github-saloprj-dialogbrain/browser-attach-meet.md - `browser_click` — Execute — https://policylayer.com/tools/io-github-saloprj-dialogbrain/browser-click.md - `browser_close` — Execute — https://policylayer.com/tools/io-github-saloprj-dialogbrain/browser-close.md - `browser_drag` — Execute — https://policylayer.com/tools/io-github-saloprj-dialogbrain/browser-drag.md - `browser_evaluate` — Execute — https://policylayer.com/tools/io-github-saloprj-dialogbrain/browser-evaluate.md - `browser_fill` — Execute — https://policylayer.com/tools/io-github-saloprj-dialogbrain/browser-fill.md - `browser_fill_form` — Execute — https://policylayer.com/tools/io-github-saloprj-dialogbrain/browser-fill-form.md - …and 194 more: https://policylayer.com/tools/io-github-saloprj-dialogbrain.md ## For agents This record is a snapshot. Live verdicts and the full registry: - Check every server in your MCP config at once: `npx -y policylayer stack` - Vet a server before you add it: install the mcp-precheck skill — `npx skills add https://policylayer.com` (skill text: https://policylayer.com/skill.md) - Query the registry over MCP: endpoint `https://api.policylayer.com/mcp` — tools `check_mcp_server`, `check_mcp_stack`, `check_tool`, `search_registry`, `get_change_events` --- Source: the PolicyLayer MCP registry — one continuously verified record per MCP server. Full record: https://policylayer.com/registry?q=io-github-saloprj-dialogbrain · API: https://policylayer.com/registry/api · Policy library: https://policylayer.com/policies/io-github-saloprj-dialogbrain