Loading ToolBench
MCP server for Google Flow (labs.google/fx/tools/flow) via browser automation
Server has 14 tools with explicit schemas and descriptions, but quality is inconsistent. Tool names lack clear verb-noun structure (flow_* prefix obscures action intent). Descriptions vary widely: some are detailed (flow_generate_image, flow_generate_video with credit warnings), others are minimal (flow_disconnect, flow_account_check). Parameter descriptions are present but often lack constraints, ranges, or format guidance. Output schemas are not documented, responses are inferred from code inspection. Error handling is minimal; no recovery guidance or actionable error messages visible. Security concerns: no evidence of secret injection for credentials; browser automation exposes sensitive operations without confirmation patterns.
Verify the logged-in Google account matches the configured expected email (Profile 3).
Launch Chrome with the configured Google profile, connect CDP, navigate to Google Flow, and verify account.
Create a new character in Google Flow Characters with name and description.
Create a new scene in Google Flow Scenes with characters and prompt.
Close the browser and clean up the MCP connection to Google Flow.
Navigate to a Google Flow page and discover all interactive elements (buttons, inputs, links, headings). Updates the internal selectors map for robust automation.
Tool names use flow_* prefix but lack clear action verbs. 'flow_disconnect' vs 'close_browser', 'flow_account_check' vs 'verify_account', LLMs cannot infer intent from prefix alone. Verb-noun naming (get_, create_, verify_) is standard in 90% of A+ tools.
Output schemas are not documented. Code inspection reveals tools return objects with fields like 'status', 'url', 'elements', 'screenshot', but no formal schema is declared. LLMs cannot plan downstream calls or extract fields reliably without documented return types.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 54 | <=2025-11-25 | v2 |
Download the most recently generated file from Google Flow.
⚠️ CES IMAGES CONSOMMENT DES CRÉDITS. Par défaut (auto_confirm=false): remplit le prompt, sélectionne le modèle/ratio, prend un screenshot et retourne "ready_for_confirmation". NE clique PAS sur Generate. Quand auto_confirm=true: vérifie d'abord que l'interface est bien en mode IMAGE (pas Vidéo), que le modèle est un modèle image, prend un screenshot de vérification, PUIS clique Generate, attend les images et les télécharge. NAN/BANANA modèles image seulement.
⚠️ CONSUMA CREDITI FLOW (Veo/Omni). Con auto_confirm=false (default): prepara il prompt video e si ferma senza generare (nessun credito). Con auto_confirm=true: invia, attende il render (minuti) e scarica l'mp4. Modello economico per test: "lite" (Veo 3.1 Lite).
Import a character from a saved JSON file into Google Flow.
Open the Google Flow Characters page and list existing characters.
Open the Google Flow Tools Gallery and list available tools.
Check current connection status: browser connected, Flow page loaded, account verified, job queue state.
Open Grid Architect in Google Flow, fill theme prompt, shot prompts, engine, ratio, and visual logic settings. Supports batch shot generation for brand campaigns.
Irreversible operations (flow_generate_image, flow_generate_video, flow_use_grid_architect) lack confirmation/dry-run patterns. auto_confirm flag exists but no explicit confirmation step or recovery guidance. Agents can accidentally consume credits without safeguards.
Parameter descriptions lack constraints and format guidance. 'model' accepts 'Nano Banana Pro', 'Nano Banana 2', 'Imagen 4' but no enum constraint. 'ratio' accepts '1:1', '16:9', etc. but format not formally specified. LLMs guess at valid values.
Error handling is minimal. No visible recovery guidance, error classification, or actionable messages. Code uses generic try-catch without structured error responses. LLMs cannot self-correct or decide whether to retry.