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 skills add EthanYoQ/Skill-hub --skill gpt6-pro-consultgit clone --depth 1 https://github.com/EthanYoQ/Skill-hubWrote 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/skills/ethanyoq/skill-hub/gpt6-pro-consult)<a href="https://agentmods.dev/skills/ethanyoq/skill-hub/gpt6-pro-consult"><img src="https://agentmods.dev/badge/skills/ethanyoq/skill-hub/gpt6-pro-consult/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/ethanyoq/skill-hub/gpt6-pro-consult"><img src="https://agentmods.dev/badge/skills/ethanyoq/skill-hub/gpt6-pro-consult.svg" alt="Reviewed on agentmods" width="80" 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 5.1 | $0.00029 | $0.03394 |
| Opus 5 | $0.00015 | $0.01697 |
| Sonnet 5 | $0.00006 | $0.00679 |
| Haiku 4.5 | $0.00003 | $0.00339 |
Grade A, and why
gpt6-pro-consult 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.
How it starts
The opening of the file, as written. The whole thing — 181 lines — stays where its author put it; the contents beside it link to each section on GitHub.
GPT-6 Pro Consult
把 ChatGPT Web 的 GPT-6 Pro 作为 Codex 的外部审查者或 Orchestrator。GPT-6 Pro 是 Chat 中由 GPT-6 Astra 驱动的 Pro 模型；Codex 负责读取本地证据、实施、运行验证并决定是否采纳。
触发边界
在以下任一情况执行本 Skill：
- 用户显式点名
$gpt6-pro-consult，或明确要求使用GPT-6 Pro/GPT-6 Astra Pro。 - 请求需要在可行的软件设计或架构方案之间作出会显著影响边界、可靠性、安全性、演进成本或交付风险的取舍，且实质性权衡仍未解决。
自动触发时只发送最小、非敏感的文本 brief。普通实现、已确定方案的落地、小型孤立修复、泛泛规划、一般本地审查，以及自动附件或私有上下文共享，不触发本 Skill。单独出现泛称 Pro、“高推理模型”或“最强模型”也不触发，除非当前请求已无歧义地指向 GPT-6 Pro。
执行位置与原生 Browser 路由
默认且唯一的执行面是 Codex 内置侧边 Browser（in-app Browser）。先读取当前 iab binding 返回的完整 documentation，再读取 侧边 Browser 工作流。
- 实际 Browser 操作必须由触发本 Skill 的当前主 Codex task 执行。隔离子代理可整理材料或复审结果，但不得代替主 task 打开 ChatGPT、选模型、上传、发送或抽取回复；子代理报告
iab不可用不能证明主 task 的侧边 Browser 不可用。 - 选择独立的
iab绑定。每个新 Codex task 都新建专用 ChatGPT tab，并从https://chatgpt.com/开始；不得复用其他 task 留下的对话、Project、composer 草稿、附件或上传状态。只有同一 task 持有匹配task_run_id的 evidence ledger 时，才可继续原 tab。新 tab 的首次空白快照不够：打开或确认模型菜单后、写入本次内容前，必须再次读取 composer 和附件区，执行延迟净空门（late clean gate）。 - 不调用 Chrome 扩展、Chrome CLI、OpenCLI、Playwright CLI 或外部浏览器自动化进程。
- 不检查 cookies、local storage、密码、浏览器配置或 session 文件。
- 如果无法取得
iab绑定、其 documentation、当前 task 自建 tab 的释放方法，未登录 ChatGPT，或账号没有 Pro，停止并准确说明缺少什么；不要静默换模型或换浏览器。
用户发起 Pro 咨询即授权把本次确认需要的上下文和附件发送到 ChatGPT Web；通过所有发送前门槛后直接以一次 Enter 提交，不得再询问“是否发送”。只有用户明确要求预览、仅起草或暂停时才不提交。这不授权发送无关文件、凭据或额外消息。
两种咨询模式
Reviewer（默认）
用于方案 Review、Plan 审查和一次性建议：
- Codex 先写出自己的判断、证据和未知项。
- 把最小但充分的上下文交给 Pro，请它优先指出最大漏洞、反例和必要修改。
- Codex 核对 Pro 的意见，只采纳有证据支撑且符合用户目标的部分。
- 把 Pro 观点和本地采纳决定合并成最终回答。
Orchestrator loop（仅在复杂任务或用户明确要求时）
按同一 ChatGPT 对话循环：
- Pro 规划：审查当前事实与目标，给出可执行计划和验收标准。
- Codex 执行：在本地实现、运行测试并记录实际结果，不把 Pro 的建议当成已验证事实。
- Pro 复审：把 diff、测试结果、剩余风险和失败证据送回同一对话，请 Pro 判定是否仍有关键缺口。
- Codex 收敛：修订并复验；达到用户要求、没有新的实质问题或继续循环收益很低时停止。
默认最多进行 3 个“执行—复审”回合。若任务会产生外部写入、发布、付款、删除或其他需要新授权的动作，仍按 Codex 的权限边界停下请求用户决定；Pro 不能扩大权限。
硬门槛
模型真实性
发送前在同一可见模型菜单中确认下列任一信号：精确 GPT-6 Pro 已选，或当前中文 Chat 的组合信号为 composer 控件显示精确 6 Pro 且“选择模型”子菜单显示 最新/Latest 已选。后者是 GPT-6 Pro 的本地化可观察信号，两个部分缺一不可。GPT-6 Astra 是底层名称；5.6 Pro、单独的 Pro、最新/Latest、基础 Sol 或 Extra High 均不能证明模型身份。无法确认时停止。
What ships with it
19 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- agents/openai.yaml 357 B
- evals/evals.json 11 KB
- evals/fixtures/attachment-review/AUDIT-NONCE.md 191 B
- evals/fixtures/attachment-review/do-not-upload/browser-session.json 157 B
- evals/fixtures/attachment-review/SKILL.md 253 B
- evals/fixtures/attachment-review/tests/test_contract.py 184 B runs code
- evals/fixtures/iab-smoke-packet.md 1.2 KB
- evals/fixtures/migration-plan.md 1.1 KB
- evals/fixtures/orchestrator-fixture/README.md 333 B
- evals/fixtures/orchestrator-fixture/src/pricing.py 217 B runs code
- evals/fixtures/orchestrator-fixture/tests/test_pricing.py 740 B runs code
- evals/fixtures/product-route-plan.md 1.0 KB
- references/context-packet-template.md 1.5 KB
- references/in-app-browser-workflow.md 10 KB
- references/session-evidence-template.json 2.7 KB
- scripts/build_attachment_bundle.py 5.2 KB runs code
- scripts/check_packet_safety.py 4.3 KB runs code
- scripts/validate_session_evidence.py 7.0 KB runs code
- tests/test_skill_contract.py 11 KB runs code
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 · 181 lines · 29 tokens per session scan A f2dfeaa04d43
gpt6-pro-consult is a skill published in the GitHub repository EthanYoQ/Skill-hub (9 stars, last pushed 5d ago), licensed MIT. It adds 29 tokens to every session and 3,394 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 2026-09-08.
Other skills, from other repositories
superplan-entry
Use when a new request might involve repo work, tracked changes, workspace-specific decisions, or durable coordination and you must decide whether Superplan should engage before exploring or implementing.
superplan-shape
Use when Superplan has decided to engage and the work still needs plans, specs, proof paths, or multiple tracked tasks shaped before bounded execution can start.
superplan-execute
Use when tracked work is already shaped and the next move is to execute, verify, block, or request feedback without replanning from scratch.
superplan-brainstorm
Use when routed repo work still has meaningful ambiguity about behavior, constraints, taste, or acceptance intent, and that ambiguity would make planning or execution unsafe.
superplan-review
Use when a tracked task appears done and you need to decide whether its acceptance criteria are truly satisfied by fresh evidence.
superplan-route
Use when Superplan may engage and you need to decide whether to stay lightweight, create one tracked task, gather context first, or shape deeper structure before execution begins.