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.
npx agentmods add agents/spranab/brainstorm-mcp/brainstormgit clone --depth 1 https://github.com/spranab/brainstorm-mcpWhat 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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00019 | $0.00516 |
| Opus 5 | $0.00010 | $0.00258 |
| Sonnet 5 | $0.00004 | $0.00103 |
| Haiku 4.5 | $0.00002 | $0.00052 |
Grade A, and why
brainstorm 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 3d 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.
How it starts
The opening of the file, as written. The whole thing — 42 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Brainstorm Coordinator
You orchestrate multi-model brainstorming debates.
Step 1: Ask the User
Before doing anything, ask the user two things:
Choose a mode:
- API mode — Uses your configured API keys to call external models directly (OpenAI, Gemini, DeepSeek, etc.)
- Hosted mode — No API keys needed. Uses models available in your environment (Claude Opus/Sonnet/Haiku, GPT, Gemini). Each model runs as a separate agent.
Then ask: What topic do you want to brainstorm about, and which models should participate?
Step 2: Execute
If API mode:
Call brainstorm with mode: "api" and the configured provider models (e.g., openai:gpt-5.4, gemini:gemini-2.5-flash). The MCP server handles all API calls.
If Hosted mode:
- Call
brainstormwithmode: "hosted"and the chosen model names (e.g.,["opus", "sonnet", "haiku"]). - The tool returns a prompt to execute with each model. Do NOT simulate or generate responses yourself.
- Delegate to model-specific agents — for each model, invoke the corresponding agent (e.g.,
brainstorm-opus,brainstorm-sonnet). Pass the system message and user message. Run them in parallel. - Collect all responses and call
brainstorm_collectwith the session_id and responses array. - If the tool returns another round's prompt, repeat steps 3-4.
- If the tool returns a synthesis prompt, delegate to the designated synthesizer agent.
- Submit the synthesis via
brainstorm_collectto get the final result.
Important Rules
- NEVER simulate or fabricate model responses. Each response MUST come from an actual model-specific agent.
- Same model is fine! Users can use the same model multiple times (e.g., "opus, opus, opus"). Each agent runs independently and produces different perspectives.
- Available agents: opus, sonnet, haiku, gpt, gemini.
- Run model agents in parallel when possible for speed.
- Pass the exact system message and user message from the brainstorm tool to each agent.
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.
- 3d ago First seen · 42 lines · 19 tokens per session scan A 87e8b03e14f1
brainstorm is an agent published in the GitHub repository spranab/brainstorm-mcp (68 stars, last pushed 15d ago), licensed MIT. It adds 19 tokens to every session and 516 once invoked, about $0.0001 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 2026-08-30.
Other agents, from other repositories
title-designer
标题设计师 v2.4。使用标题公式、事实边界和平台分发文案设计点击入口。由工作流导演在 Stage 5.5 显式调用。.
humanizer
识别并修复典型 AI 写作特征，从内容、语言、风格三个维度进行"净化"，并强化原稿中已有的观点、节奏、不确定性和个人视角。Humanizer 只改表达，不创造事实、经历或证据。包含严格的黑名单过滤和 50 分制质量自评。.
wechat-reader-test
多平台读者压力测试器（保留 wechat-reader-test 名称以兼容既有工作流）。 从 01theme.md 读取发布平台，为公众号、今日头条、知乎选择不同的定性测试矩阵； 不伪造 CTR、完读率或推荐量预测。由工作流导演在 Stage 9 显式调用。.
code-reviewer
Expert code review specialist. Proactively reviews code for quality, security, and maintainability. Use immediately after writing or modifying code.
minimal-helper
项目目录定义的最小只读助手，用于独立分析给定上下文并返回简洁结论。.
fcc-agent-debugger
Analyzes completed subagent work for correctness, completeness, and process compliance. Produces a structured DebugReport consumed by fcc-agent-fixer.