360 Secure SkillHub
360 安全技能中心

DualAgentDebate

运行脚本在 ChatGPT 与从 Open Brain 检索到的既有想法之间进行最多 3 轮结构化辩论，按语义相似度判断是否达成一致，并把最终结论写回 Open Brain 记忆库，用于在决策前把想法与过往思考做对比和压力测试。

tvaloki
tvaloki
热度 54.96
最近更新 2026-07-16 22:09:39
分类：知识依赖：代码执行 / MCP 服务 / 第三方 API / 内部调用大模型应用场景：研究 / 笔记整理 / 工作流

安装技能

复制下面的话术发给你的 Agent，它会通过 secure-skills 检索并安装该技能。

执行命令安装此 skill： npx secure-skills-cli add dual-agent-debate
Metadata
name
dual-agent-debate
description
Run a structured 2-agent debate loop between ChatGPT (OpenAI API) and the user's own thoughts retrieved from Supabase Open Brain via MCP, iterating up to 3 rounds (or until semantic agreement), then write the final outcome back to Open Brain. Use when user asks to compare/pressure-test an idea against prior thoughts or memory.

DualAgentDebate

Execute scripts/dual_agent_debate.py to run the debate loop.

Setup

Set required environment variables:

export OPENBRAIN_MCP_URL="http://127.0.0.1:54321/mcp"
# optional if MCP is protected
export OPENBRAIN_MCP_TOKEN="..."

Optional: if you want direct OpenAI API mode, set:

export OPENAI_API_KEY="..."

If OPENAI_API_KEY is not set, the script uses openclaw agent (OAuth-backed local setup) for debate responses.

Optional tool/model overrides (defaults shown):

export OPENBRAIN_CONTEXT_TOOL="search_docs"
export OPENBRAIN_SQL_TOOL="execute_sql"
export DEBATE_MODEL="gpt-4o-mini"

Run

python3 skills/dual-agent-debate/scripts/dual_agent_debate.py \
  --query "Should I migrate this service to Supabase edge functions?"

Optional explicit thoughts:

python3 skills/dual-agent-debate/scripts/dual_agent_debate.py \
  --query "Should we launch this feature now?" \
  --thoughts "My concern is reliability and on-call burden." \
  --rounds 3 \
  --agreement-threshold 0.9

Behavior

  1. Pull context from Open Brain MCP (OPENBRAIN_CONTEXT_TOOL).
  2. Pull related prior thoughts from public.thoughts via MCP execute_sql unless --thoughts is provided.
  3. Ask ChatGPT to debate the query using context and thoughts.
  4. Compute semantic similarity (OpenAI embeddings) between ChatGPT reply and thoughts.
  5. Repeat up to 3 rounds or stop early on agreement threshold.
  6. Persist full outcome into public.memories via MCP execute_sql.

Notes

  • If your Open Brain MCP uses different tool names or argument schemas, set the tool env vars accordingly.
  • The script uses MCP tools/call JSON-RPC shape; point OPENBRAIN_MCP_URL at your MCP HTTP endpoint.