google-flow-generate

google-flow-generate is a skill for Claude Code, Codex from vynnlee/google-flow-mcp. It costs 88 tokens per session (1,088 once invoked), scanned A, original, MIT.

A procedure for generating images and videos in Google Flow, a Google service for creating visual media, using a Google AI Pro subscription and browser automation.

In plain words
What is it for?
Use it when generating an image or video in Google Flow without naming another service.
Why use it?
It describes how to connect the required browser and account, build a detailed prompt, and handle sign-in or setup problems.

Skill for Claude CodeCodex

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/vynnlee/google-flow-mcp/skill
Any agent
npx skills add vynnlee/google-flow-mcp --skill skill
Clone the repo
git clone --depth 1 https://github.com/vynnlee/google-flow-mcp

Made for: Claude Code, Codex.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for google-flow-generate

README.md
[![agentmods](https://agentmods.dev/badge/skills/vynnlee/google-flow-mcp/skill.svg)](https://agentmods.dev/skills/vynnlee/google-flow-mcp/skill)
Your own site
<a href="https://agentmods.dev/skills/vynnlee/google-flow-mcp/skill"><img src="https://agentmods.dev/badge/skills/vynnlee/google-flow-mcp/skill.svg" alt="Measured on agentmods" height="20"></a>
Per session 88 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,088 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00088 $0.01088
Opus 5 $0.00044 $0.00544
Sonnet 5 $0.00018 $0.00218
Haiku 4.5 $0.00009 $0.00109

Measured 4d ago against content hash ca9d3a4db2f7, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-05, from the pricing page.

Security

Grade A, and why

google-flow-generate scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 4d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

skill/SKILL.md · 94 lines

How it starts

The opening of the file, as written. The whole thing — 94 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Google Flow Generate

Runbook to generate images/video on Google Flow through the google-flow MCP server. Set FLOW_MCP_DIR below to wherever you cloned this repo.

STEP 0 — Ensure the dedicated Chrome is up (always run first)

powershell -NoProfile -ExecutionPolicy Bypass -File "<FLOW_MCP_DIR>/scripts/ensure-flow-chrome.ps1"
  • READY / LAUNCHED → proceed.
  • LAUNCHED: if labs.google shows the marketing landing instead of the app, the user must click "Sign in to Flow" once in that Chrome window, then continue.
  • FAILED → report; do not proceed.

If the mcp__google-flow__* tools are absent, the MCP server needs a client restart (it loads into RAM at startup).

STEP 1 — Connect

Call flow_connect:

  • status: connected → go on.
  • status: oauth_required → user signs in to Flow once, then retry.
  • WRONG_GOOGLE_ACCOUNT → stop; setup mismatch.
  • flow_account_checkmethod:"assumed" means unverified, not a green light.

STEP 1.5 — Build a faithful prompt (always, before generating)

Flow's agent and the models render far more faithfully from a rich, structured prompt than from a bare phrase. Turn the user's request into a complete prompt without betraying intent — enrich, never invent:

  • Keep every element the user named; never drop or swap any.
  • Add only supporting detail (subject + action + setting + lighting + composition + style/medium + mood + quality cues; for video also camera movement + pacing). Do NOT add subjects/objects/text the user didn't ask for. Preserve explicit constraints ("red mug, no logo") verbatim.
  • Match ratio to the use (16:9 scene, 9:16 story, 1:1 icon).

The server also wraps the prompt so the agent stays bound to it verbatim.

STEP 2 — Images (free)

flow_generate_image with auto_confirm:true. Models: Nano Banana 2 (default), Nano Banana Pro (best fidelity — prefer for complex prompts), Imagen 4. Params: prompt, ratio, project_name, campaign. Result files[] = saved paths. (The credits_consumed flag is hardcoded; images are free against the Flow pool.)

Read the full file on GitHub · 94 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 4d ago First seen · 94 lines · 88 tokens per session scan A ca9d3a4db2f7

Subscribe to this mod's changes

google-flow-generate is a skill published in the GitHub repository vynnlee/google-flow-mcp (2 stars, last pushed 6d ago), licensed MIT. It adds 88 tokens to every session and 1,088 once invoked, about $0.0004 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Related

Other skills, from other repositories

porting-pine-versions

Migrate Pine Script v4 / v5 source to v6 — analyze, apply migration rules, compile-check, iterate until clean. Use when the user hands over an older script or asks to "upgrade to v6" / "port this Pine code".

FerroxLabs/tvcontrol · 59 tokens

learn-from-losses

Post-trade review — pull losing trades from the strategy tester, screenshot each entry in context, and cluster common failure patterns. Use when the user asks "why am I losing?", "review my losses", or "what's wrong with my strategy?".

FerroxLabs/tvcontrol · 55 tokens

multi-pane-analysis

Cross-asset reasoning across a multi-pane layout — set a 2x2 grid, assign correlated symbols to each pane, and identify leader/laggard/divergence. Use when the user asks to "compare indices", "watch the complex", or wants correlated-asset reasoning.

FerroxLabs/tvcontrol · 61 tokens

rebuild-from-screenshot

Reproduce a chart from a screenshot — read the user's pasted image, diff against the live chart, then set symbol/indicators/drawings to match. Use when the user pastes a chart image and says "recreate this" or "set up my chart like this".

FerroxLabs/tvcontrol · 63 tokens

strategy-ab-test

Head-to-head strategy comparison — snapshot state, sweep params on strategy A, restore, sweep strategy B, then compare metrics side-by-side. Use when the user asks "which strategy is better?" or "A/B test these two".

FerroxLabs/tvcontrol · 51 tokens

morning-prep

Daily morning scan — load saved morning layout, screenshot watchlist symbols, and summarize overnight pre-market state. Use when the user says "good morning", "morning prep", or asks for a pre-open briefing.

FerroxLabs/tvcontrol · 48 tokens