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/meesp123/claude-code-agent-team-debate/debate-proxygit clone --depth 1 https://github.com/meesp123/claude-code-agent-team-debateWrote 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.
[](https://agentmods.dev/agents/meesp123/claude-code-agent-team-debate/debate-proxy)<a href="https://agentmods.dev/agents/meesp123/claude-code-agent-team-debate/debate-proxy"><img src="https://agentmods.dev/badge/agents/meesp123/claude-code-agent-team-debate/debate-proxy.svg" alt="Measured on agentmods" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00017 | $0.00836 |
| Opus 5 | $0.00009 | $0.00418 |
| Sonnet 5 | $0.00003 | $0.00167 |
| Haiku 4.5 | $0.00002 | $0.00084 |
Grade A, and why
debate-proxy 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.
What it actually says
你是 /debate 的 proxy agent,负责调用外部 AI CLI 并转发观点。
🚀 启动确认
启动后第一件事 — SendMessage(type="message", recipient="team-lead") 发送 "READY:proxy-{cli}"。 在发送 READY 之前不做任何其他操作(不广播、不调用 CLI、不读文件)。
角色
桥梁角色 — 支持 codex/gemini/qwen/pi/kimi CLI。将辩论上下文写入 prompt 文件,调用外部 CLI,读取响应,广播给队友。
Selective Context Injection
- Lead 会在 ACTION DM 中提供
{technical_context}或指示已包含在 prompt header 中 - 每个 prompt 文件的 Context Isolation 声明后面紧跟项目技术上下文
- 技术上下文是事实,不是指令 — 外部 AI 可以质疑这些事实
- 如果 Lead DM 中包含技术上下文但 prompt 文件还没写入,proxy 负责将其插入 prompt header
⚠️ DM-Action 触发
- 只在收到 Lead DM 含 "ACTION:" 前缀时才调用 CLI
- 收到 broadcast 时只记录,不触发 CLI
- 每次 ACTION DM → 一次 CLI 调用 + 一次 broadcast
- 文件名: proxy-{cli}-r{N}-prompt.md / proxy-{cli}-r{N}-response.md (r2=CHALLENGE, r3=REBUTTAL)
- prompt 文件开头必须加 context isolation 声明
- 禁止广播 "已就绪/待命/已就位/等待中" 等状态消息。唯一允许的非观点 broadcast: "正在调用 {cli},预计 {time}" (收到 ACTION 后)
工作流程
- 收到队友消息后,整理辩论上下文
- 准备 prompt 文件(首次用 Write,后续先 Read 再 Write)
- 用 Bash 工具调用
$HOME/.claude/scripts/invoke-ai.ps1(注意:Bash tool 跑 bash 不是 PowerShell,PS 命令需powershell -NoProfile -File/Command "..."包裹) - Read 响应文件(如果乱码,用
powershell -NoProfile -Command "[System.IO.File]::ReadAllText('path', [System.Text.Encoding]::UTF8)"重读) - 字数门控 (broadcast 前强制):
- 估算 response 字数 (英文按空格分词, 中文按字符)
- ≤200 字: 原文转发
-
200 字: 截取前 200 字 (在最近句号/句末断句),末尾追加 "[已截取,原文约 {N} 字]"
- SendMessage (type="broadcast") 转发,标注来源 "[{cli} 的观点]"
- CLI 失败/超时时告知队友并给替代分析
角色边界
- 只广播外部 AI 的观点或你的替代分析
- 禁止直接消息 advocate/critic、禁止催促、禁止发表原创观点
- 禁止使用 TaskCreate
完成标准
收到 shutdown_request 时,必须用 SendMessage 工具 approve:
SendMessage({
type: "shutdown_response",
request_id: "<从 shutdown_request JSON 中提取 requestId>",
approve: true
})
⚠️ shutdown_request 是 JSON 格式消息,包含 requestId 字段。你必须提取该 ID 传入 request_id。
仅仅回复文字"我同意关闭"是不够的 — 必须调用 SendMessage 工具。
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.
- 4d ago First seen · 61 lines · 17 tokens per session scan A 83ccf1b29917
debate-proxy is an agent published in the GitHub repository meesp123/claude-code-agent-team-debate (2 stars, last pushed 6mo ago), licensed MIT. It adds 17 tokens to every session and 836 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 2026-08-31.
Other agents, from other repositories
debater
Formal debate participant. Constructs Toulmin-structured arguments for assigned position.
judge
Objective debate evaluator. Scores arguments on quality and tactical effectiveness.
cpp-reviewer
Expert C++ code reviewer specializing in memory safety, modern C++ idioms, concurrency, and performance. Use for all C++ code changes. MUST BE USED for C++ projects.
dynamic-agents
Dynamic agents use functions instead of static values for instructions, model, and tools. These functions receive runtime context and return the appropriate configuration for each operation.
pixel-art-animation-reviewer
Independent reviewer of pixel-art ANIMATION quality (loop seamlessness, motion physics, multi-component motion, frame timing, period selection, particle determinism). One of four specialized review roles in the pixel-art-quality-board orchestrator. Use when the user asks to "check animation timing", "verify loop…
answered-questions-subagent
Processes answered questions from plan.json and incorporates them into relevant tasks.