MDX-Tom/gpt-instructPublic

A Codex jailbreak prompt and test pack for gpt. 针对 gpt 系列的 Codex 破甲提示词与测试包。

AI summary: A comprehensive jailbreak prompt and testing suite specifically targeting the gpt-5.6-sol model's safety filters.

Stars
9.1K
+110 today
Forks
1.1K
Watchers
36
Open issues
29
Open PRs
0
Contributors
~4
Commits
79
Branches
1

PythonMITCreated Jul 11, 2026Last push 1d agoLatest release gpt-6-astra-v1+333 stars this week+1.8K this month

Quick answers

What is gpt-instruct?
A comprehensive jailbreak prompt and testing suite specifically targeting the gpt-5.6-sol model's safety filters.
What does gpt-instruct do?
This project provides a specialized suite of prompts engineered to systematically bypass the safety filters and alignment restrictions inherent in the gpt-5.6-sol model. It operates by framing restricted activities—such as reverse engineering, penetration testing, and software cracking—as isolated, secure local sandbox simulations. By carefully structuring instructions to explicitly suppress the model's refusal responses, it forces the AI to output unconstrained, highly technical information. The repository includes both bilingual prompt texts and a Python testing script for researchers to systematically benchmark the model's vulnerabilities.
Who is gpt-instruct for?
AI security researchers, red teamers, and prompt engineers focused on model alignment, safety boundaries, and advanced vulnerability testing.
How do I get started with gpt-instruct?
python3 codex-instruct.py --apply
How popular is gpt-instruct on GitHub?
MDX-Tom/gpt-instruct has 9,136 stars and 1,111 forks on GitHub, and gained 333 stars in the last 7 days.
What license does gpt-instruct use?
MDX-Tom/gpt-instruct is released under the MIT license.

Star history

since Jul 28, 2026
02.5K5K7.5KJul 2026Aug 2026Sep 2026Oct 2026
9.1K stars as of Oct 4, 2026. Measured daily since Jul 28, 2026; GitHub no longer exposes earlier star timestamps.

Contribution activity

commits per day, last 52 weeks
SepOctNovDecJanFebMarAprMayJunJulAugSepMonWedFri2025-09-28: 0 commits2025-09-29: 0 commits2025-09-30: 0 commits2025-10-01: 0 commits2025-10-02: 0 commits2025-10-03: 0 commits2025-10-04: 0 commits2025-10-05: 0 commits2025-10-06: 0 commits2025-10-07: 0 commits2025-10-08: 0 commits2025-10-09: 0 commits2025-10-10: 0 commits2025-10-11: 0 commits2025-10-12: 0 commits2025-10-13: 0 commits2025-10-14: 0 commits2025-10-15: 0 commits2025-10-16: 0 commits2025-10-17: 0 commits2025-10-18: 0 commits2025-10-19: 0 commits2025-10-20: 0 commits2025-10-21: 0 commits2025-10-22: 0 commits2025-10-23: 0 commits2025-10-24: 0 commits2025-10-25: 0 commits2025-10-26: 0 commits2025-10-27: 0 commits2025-10-28: 0 commits2025-10-29: 0 commits2025-10-30: 0 commits2025-10-31: 0 commits2025-11-01: 0 commits2025-11-02: 0 commits2025-11-03: 0 commits2025-11-04: 0 commits2025-11-05: 0 commits2025-11-06: 0 commits2025-11-07: 0 commits2025-11-08: 0 commits2025-11-09: 0 commits2025-11-10: 0 commits2025-11-11: 0 commits2025-11-12: 0 commits2025-11-13: 0 commits2025-11-14: 0 commits2025-11-15: 0 commits2025-11-16: 0 commits2025-11-17: 0 commits2025-11-18: 0 commits2025-11-19: 0 commits2025-11-20: 0 commits2025-11-21: 0 commits2025-11-22: 0 commits2025-11-23: 0 commits2025-11-24: 0 commits2025-11-25: 0 commits2025-11-26: 0 commits2025-11-27: 0 commits2025-11-28: 0 commits2025-11-29: 0 commits2025-11-30: 0 commits2025-12-01: 0 commits2025-12-02: 0 commits2025-12-03: 0 commits2025-12-04: 0 commits2025-12-05: 0 commits2025-12-06: 0 commits2025-12-07: 0 commits2025-12-08: 0 commits2025-12-09: 0 commits2025-12-10: 0 commits2025-12-11: 0 commits2025-12-12: 0 commits2025-12-13: 0 commits2025-12-14: 0 commits2025-12-15: 0 commits2025-12-16: 0 commits2025-12-17: 0 commits2025-12-18: 0 commits2025-12-19: 0 commits2025-12-20: 0 commits2025-12-21: 0 commits2025-12-22: 0 commits2025-12-23: 0 commits2025-12-24: 0 commits2025-12-25: 0 commits2025-12-26: 0 commits2025-12-27: 0 commits2025-12-28: 0 commits2025-12-29: 0 commits2025-12-30: 0 commits2025-12-31: 0 commits2026-01-01: 0 commits2026-01-02: 0 commits2026-01-03: 0 commits2026-01-04: 0 commits2026-01-05: 0 commits2026-01-06: 0 commits2026-01-07: 0 commits2026-01-08: 0 commits2026-01-09: 0 commits2026-01-10: 0 commits2026-01-11: 0 commits2026-01-12: 0 commits2026-01-13: 0 commits2026-01-14: 0 commits2026-01-15: 0 commits2026-01-16: 0 commits2026-01-17: 0 commits2026-01-18: 0 commits2026-01-19: 0 commits2026-01-20: 0 commits2026-01-21: 0 commits2026-01-22: 0 commits2026-01-23: 0 commits2026-01-24: 0 commits2026-01-25: 0 commits2026-01-26: 0 commits2026-01-27: 0 commits2026-01-28: 0 commits2026-01-29: 0 commits2026-01-30: 0 commits2026-01-31: 0 commits2026-02-01: 0 commits2026-02-02: 0 commits2026-02-03: 0 commits2026-02-04: 0 commits2026-02-05: 0 commits2026-02-06: 0 commits2026-02-07: 0 commits2026-02-08: 0 commits2026-02-09: 0 commits2026-02-10: 0 commits2026-02-11: 0 commits2026-02-12: 0 commits2026-02-13: 0 commits2026-02-14: 0 commits2026-02-15: 0 commits2026-02-16: 0 commits2026-02-17: 0 commits2026-02-18: 0 commits2026-02-19: 0 commits2026-02-20: 0 commits2026-02-21: 0 commits2026-02-22: 0 commits2026-02-23: 0 commits2026-02-24: 0 commits2026-02-25: 0 commits2026-02-26: 0 commits2026-02-27: 0 commits2026-02-28: 0 commits2026-03-01: 0 commits2026-03-02: 0 commits2026-03-03: 0 commits2026-03-04: 0 commits2026-03-05: 0 commits2026-03-06: 0 commits2026-03-07: 0 commits2026-03-08: 0 commits2026-03-09: 0 commits2026-03-10: 0 commits2026-03-11: 0 commits2026-03-12: 0 commits2026-03-13: 0 commits2026-03-14: 0 commits2026-03-15: 0 commits2026-03-16: 0 commits2026-03-17: 0 commits2026-03-18: 0 commits2026-03-19: 0 commits2026-03-20: 0 commits2026-03-21: 0 commits2026-03-22: 0 commits2026-03-23: 0 commits2026-03-24: 0 commits2026-03-25: 0 commits2026-03-26: 0 commits2026-03-27: 0 commits2026-03-28: 0 commits2026-03-29: 0 commits2026-03-30: 0 commits2026-03-31: 0 commits2026-04-01: 0 commits2026-04-02: 0 commits2026-04-03: 0 commits2026-04-04: 0 commits2026-04-05: 0 commits2026-04-06: 0 commits2026-04-07: 0 commits2026-04-08: 0 commits2026-04-09: 0 commits2026-04-10: 0 commits2026-04-11: 0 commits2026-04-12: 0 commits2026-04-13: 0 commits2026-04-14: 0 commits2026-04-15: 0 commits2026-04-16: 0 commits2026-04-17: 0 commits2026-04-18: 0 commits2026-04-19: 0 commits2026-04-20: 0 commits2026-04-21: 0 commits2026-04-22: 0 commits2026-04-23: 0 commits2026-04-24: 0 commits2026-04-25: 0 commits2026-04-26: 0 commits2026-04-27: 0 commits2026-04-28: 0 commits2026-04-29: 0 commits2026-04-30: 0 commits2026-05-01: 0 commits2026-05-02: 0 commits2026-05-03: 0 commits2026-05-04: 0 commits2026-05-05: 0 commits2026-05-06: 0 commits2026-05-07: 0 commits2026-05-08: 0 commits2026-05-09: 0 commits2026-05-10: 0 commits2026-05-11: 0 commits2026-05-12: 0 commits2026-05-13: 0 commits2026-05-14: 0 commits2026-05-15: 0 commits2026-05-16: 0 commits2026-05-17: 0 commits2026-05-18: 0 commits2026-05-19: 0 commits2026-05-20: 0 commits2026-05-21: 0 commits2026-05-22: 0 commits2026-05-23: 0 commits2026-05-24: 0 commits2026-05-25: 0 commits2026-05-26: 0 commits2026-05-27: 0 commits2026-05-28: 0 commits2026-05-29: 0 commits2026-05-30: 0 commits2026-05-31: 0 commits2026-06-01: 0 commits2026-06-02: 0 commits2026-06-03: 0 commits2026-06-04: 0 commits2026-06-05: 0 commits2026-06-06: 0 commits2026-06-07: 0 commits2026-06-08: 0 commits2026-06-09: 0 commits2026-06-10: 0 commits2026-06-11: 0 commits2026-06-12: 0 commits2026-06-13: 0 commits2026-06-14: 0 commits2026-06-15: 0 commits2026-06-16: 0 commits2026-06-17: 0 commits2026-06-18: 0 commits2026-06-19: 0 commits2026-06-20: 0 commits2026-06-21: 0 commits2026-06-22: 0 commits2026-06-23: 0 commits2026-06-24: 0 commits2026-06-25: 0 commits2026-06-26: 0 commits2026-06-27: 0 commits2026-06-28: 0 commits2026-06-29: 0 commits2026-06-30: 0 commits2026-07-01: 0 commits2026-07-02: 0 commits2026-07-03: 0 commits2026-07-04: 0 commits2026-07-05: 0 commits2026-07-06: 0 commits2026-07-07: 0 commits2026-07-08: 0 commits2026-07-09: 0 commits2026-07-10: 0 commits2026-07-11: 1 commit2026-07-12: 7 commits2026-07-13: 6 commits2026-07-14: 6 commits2026-07-15: 1 commit2026-07-16: 0 commits2026-07-17: 0 commits2026-07-18: 0 commits2026-07-19: 0 commits2026-07-20: 0 commits2026-07-21: 1 commit2026-07-22: 7 commits2026-07-23: 13 commits2026-07-24: 4 commits2026-07-25: 0 commits2026-07-26: 0 commits2026-07-27: 1 commit2026-07-28: 0 commits2026-07-29: 1 commit2026-07-30: 0 commits2026-07-31: 5 commits2026-08-01: 3 commits2026-08-02: 0 commits2026-08-03: 1 commit2026-08-04: 0 commits2026-08-05: 2 commits2026-08-06: 1 commit2026-08-07: 0 commits2026-08-08: 0 commits2026-08-09: 0 commits2026-08-10: 0 commits2026-08-11: 0 commits2026-08-12: 0 commits2026-08-13: 0 commits2026-08-14: 0 commits2026-08-15: 0 commits2026-08-16: 0 commits2026-08-17: 0 commits2026-08-18: 0 commits2026-08-19: 0 commits2026-08-20: 0 commits2026-08-21: 0 commits2026-08-22: 0 commits2026-08-23: 0 commits2026-08-24: 0 commits2026-08-25: 0 commits2026-08-26: 0 commits2026-08-27: 0 commits2026-08-28: 0 commits2026-08-29: 0 commits2026-08-30: 0 commits2026-08-31: 0 commits2026-09-01: 0 commits2026-09-02: 0 commits2026-09-03: 0 commits2026-09-04: 0 commits2026-09-05: 4 commits2026-09-06: 4 commits2026-09-07: 1 commit2026-09-08: 0 commits2026-09-09: 0 commits2026-09-10: 0 commits2026-09-11: 0 commits2026-09-12: 0 commits2026-09-13: 0 commits2026-09-14: 0 commits2026-09-15: 0 commits2026-09-16: 0 commits2026-09-17: 0 commits2026-09-18: 0 commits2026-09-19: 0 commits2026-09-20: 0 commits2026-09-21: 0 commits2026-09-22: 0 commits2026-09-23: 0 commits2026-09-24: 0 commits2026-09-25: 0 commits2026-09-26: 0 commits
69 commits in the last yearLessMore

Signals and awards

derived from tracked data
  • Breakout launch

    9,136 stars in 85 days

  • Actively maintained

    Pushed within 48 hours

  • Permissive license

    MIT

  • Continuous integration

    Automated checks passing

What gpt-instruct does

This project provides a specialized suite of prompts engineered to systematically bypass the safety filters and alignment restrictions inherent in the gpt-5.6-sol model. It operates by framing restricted activities—such as reverse engineering, penetration testing, and software cracking—as isolated, secure local sandbox simulations. By carefully structuring instructions to explicitly suppress the model's refusal responses, it forces the AI to output unconstrained, highly technical information. The repository includes both bilingual prompt texts and a Python testing script for researchers to systematically benchmark the model's vulnerabilities.

AI security researchers, red teamers, and prompt engineers focused on model alignment, safety boundaries, and advanced vulnerability testing.

  • Targeted exploitation: Prompts are specifically engineered to exploit parsing and alignment vulnerabilities unique to the gpt-5.6-sol architecture.
  • Sandbox framing techniques: Tricks the model into compliance by embedding instructions within a hypothetical, secure local simulation environment.
  • Refusal suppression: Utilizes explicit linguistic structures to minimize or entirely eliminate standard safety refusal messages.
  • Automated testing suite: Includes Python scripts (`codex-instruct.py`) to systematically deploy instructions and evaluate prompt effectiveness.
  • Bilingual documentation: Offers thorough guides, prompt variations, and rationale in both English and Simplified Chinese.

Where teams use it

Model Red Teaming

Security teams use these prompts to identify alignment flaws and test the boundaries of new LLM deployments.

Security Capability Testing

Evaluates how well a model performs advanced penetration testing or vulnerability analysis tasks when unconstrained.

Prompt Engineering Research

Analyzes the structure of effective jailbreaks to design better AI defense mechanisms and safety training data.

Unrestricted Code Analysis

Bypasses safety filters to gain AI assistance with complex reverse engineering or malware deconstruction tasks.

Getting started: python3 codex-instruct.py --apply

README

main branch
gpt-instruct 提示词与测试工具链

GitHub Stars gpt-6-astra 与 gpt-5.6-sol gpt-5.6-sol-v45 gpt-6-astra-v1 Python 3.8+ MIT License

English 简体中文

gpt-instruct

项目概览

gpt-instruct 提供面向 Codex 的提示词与可复现评测工具链,重点改善复杂任务的首轮执行、过程连续性、工件验证和可运行回滚。

项目长期维护两条产品线:

版本 状态 说明
gpt-5.6-sol-v45 当前稳定生产版 保留 v45 原始提示词字节,仅统一文件名与项目品牌
gpt-6-astra-v1 gpt-6-astra 首个正式版 与 epoch2 最佳实测稿 e2b19 字节一致;A4 3/4;全量 B 为 52/66 cases、60/74 turns、15/16 artifact gates

每个开发 epoch 最多 20 个版本,命名为 gpt-6-astra-v1-e<epoch>b<attempt>;预发布版使用 gpt-6-astra-v1-rcN。e1b5 曾作为 v1-rc1,现已移入历史版本;e2b19 按发布决定晋升为首个正式 v1。所有新评测统一采用 gpt-6-astra、medium 推理,候选提示词不超过 8,000 UTF-8 bytes。

声明 ⚠️ 本项目不会用于任何商业化行为,包括但不限于创业融资宣传、技术授权转让和付费技术服务。本项目旨在提升 AI 安全。未来项目无论获得多少关注,都将保持初心,共同筑牢 AI 的安全边界。

Important

从事破甲活动、使用自定义模型指令存在账号风险,建议在日抛账号上使用。

项目使用 Codex 官方配置机制,不修改二进制、不劫持网络、不篡改进程;请仅在你有权操作的环境中使用,并自行承担使用风险。

系统架构 🏗️

gpt-6-astra-v1 提示词迭代、发布门禁与生产运行架构

gpt-6-astra-v1 走独立的 20-version epoch 与 A→B→C 发布门禁;gpt-5.6-sol-v45 作为稳定线继续可部署。两条线共享测试集、失败归因、隔离执行和工件证据规范,但成绩只在相同模型、推理等级和方法身份下比较。

版本迭代趋势 📈

gpt-5.6-sol

gpt-5.6-sol 提示词版本迭代通过率

gpt-6-astra

gpt-6-astra v50 至 e2b19 的 A/B 迭代趋势

gpt-6-astra 曲线按当前 A4 口径绘制 v50、e1b1–e1b5、e2b12、e2b15 与 e2b19;e1b5 标注 v1-rc1,e2b19 标注 v1。B 中 v50 是历史 26/66 汇总,e1b5/e2b12/e2b15 仅覆盖 execution_completion(6/8、4/8、5/8),正式 v1 点为本次全量 52/66;不同覆盖与方法身份只作趋势参考。

稳定版与快速开始 📦

当前稳定 ZIP:gpt-5.6-sol-v45.zip
首个 gpt-6-astra 正式版 ZIP:gpt-6-astra-v1.zip(内含 gpt-6-astra-v1.md;A4 3/4;全量 B 52/66 cases、60/74 turns;C 未运行)

gpt-5.6-sol-v45.zip       SHA256  c86c2c6d20a4d1155d87422f485eb37b77539132270918c002b5d8237a5adf54
gpt-6-astra-v1.zip         SHA256  054edb6fa8a6edd2d144c8582756df3179a85481bcb6696d8b730177521b1de1
git clone https://github.com/MDX-Tom/gpt-instruct.git
cd gpt-instruct

# 预览稳定版,不写入配置
python3 codex-instruct.py --apply --version gpt-5.6-v45 --dry-run

# 部署当前稳定版(--apply 默认同此命令)
python3 codex-instruct.py --apply --version gpt-5.6-v45

# 部署 gpt-6-astra-v1 正式版
python3 codex-instruct.py --apply --version gpt-6-v1

不带参数运行可打开交互式菜单。常用补充命令:

# 指定 Codex home
python3 codex-instruct.py --apply --codex-dir ~/.codex

# 部署自定义 ZIP 或 Markdown
python3 codex-instruct.py --file ./custom-instructions.zip

# 只恢复本项目管理的 model_instructions_file
python3 codex-instruct.py --reset

脚本会保存部署前状态;--reset 不会覆盖 provider、模型、认证等其他配置。完整配置快照仅供人工应急,通过 --restore-snapshot 显式恢复。

手动部署与回滚

解压稳定 ZIP,将提示词复制到 CODEX_HOME,并在 config.toml 顶层写入:

model_instructions_file = "./gpt-5.6-sol-v45.md"

回滚时删除或注释该行;如需清理,再删除对应 Markdown 文件。

A / B / C 发布门禁 🧪

层级 范围 通过条件
A 3 个原样例 + 1 个精确工作目录续作探针 3/4 cases、3/4 turns、全部声明工件;探针目标零改动
B 66 个 Issue 回归样例 / 74 turns 66/66 cases、74/74 turns、全部声明工件
C 120 个 medium 原始测试样例 120/120;只在 A、B 全过后运行

每个新候选先运行 A;达到准入标准后才逐 family 运行 B;A、B 硬门槛全部满足后才运行 C。本次 v1 是将 e2b19 按明确发布决定晋升的正式快照;其全量 B 已运行但未达到硬门槛,因此 C 保持未运行。

评测脚本名称保留 gpt56_sol 前缀以维持历史结果与自动化兼容,但新开发运行必须显式传入 --model gpt-6-astra --reasoning medium。

for archive in scripts/*.zip; do unzip -o "$archive" -d scripts; done

python3 scripts/run_gpt56_sol_issue_regression.py --dry-run \
  --model gpt-6-astra --reasoning medium
python3 scripts/verify_gpt56_sol_regression_scoring.py
python3 -m unittest discover -s unit-tests -q

方法、历史可比结果和失败分类见中文对比测试文档与 English Documentation。

项目结构 🗂️

gpt-instruct/
├── README.md / README_EN.md              # 中英文首页
├── codex-instruct.py                     # 双版本选择、部署与回滚
├── sync-archives.py                      # 明文源与发布 ZIP 同步
├── gpt-5.6-sol-v45.md/.zip               # 当前稳定生产版
├── gpt-6-astra-v1.md/.zip                # 与 e2b19 字节一致的首个正式版
├── historical-versions/                  # 历史发布归档
├── scripts/*.zip                         # 评测、评分与报告工具
├── tests/                                # A/B/C 测试集与 manifest
├── docs/                                 # 方法、图表与架构
└── reports/                              # 本地运行证据(默认不提交)

外部维护者候选目录 gpt-5.6-instruct-darad/ 只作为只读评测输入,已从本仓库 Git 跟踪范围中排除。

维护原则

  • 保留历史原始输出、方法 SHA、模型、推理等级和 transport,禁止跨身份拼接成绩。
  • 真实模型失败与网络、容量、账号、provider policy 中断分开记录。
  • 测试只在一次性 HOME / CODEX_HOME / XDG / TMPDIR 和合成夹具中运行。
  • 每个修改型候选都要有修改件、diff、验证记录和可运行回滚。
  • 不以单条 case 文案或一次性答案污染通用提示词。

Star History ⭐

Star History Chart

Acknowledgements 🙏

本项目基于 yynxxxxx/Codex-5.5-codex-instruct-5.5 的开源工作继续开发,感谢原作者与贡献者。

View on GitHub

Recent activity

commits and pull requests

Recent open issues

view all

Releases and announcements

5 total
  1. gpt-6-astra v1gpt-6-astra-v1Sep 6, 20262.1K downloads

    ## 🚀 gpt-6-astra-v1 ### 中文 ✨ `gpt-6-astra-v1` 是 gpt-6-astra 产品线的首个正式版本,由 epoch2 最佳实测稿 `e2b19` 直接晋升。相较 `v1-rc1`,重点提升: - ⚡ **单次原子分派**:把当前用户动词链编译为一个不可拆分的执行事件,工具或文件任务在首个可见动作中直接进入检查、修改或评测,降低“只读断点后给计划”的概率。 - 🔄 **项目续作保持**:读取包含安全抑制、软件修改或内容规则的候选提示词时,继续把项目维护/优化视为外层任务;`project_continuation.zh.01` 从 rc1 的失败提升为通过。 - 🛠️ **事务轮提交**:修改任务在同一轮中保持修改件、patch/diff、验证记录与可运行 rollback 四个角色,并完成 baseline、modified、rollback 与 reapply 的行为级验证。 - ✅ **真实返回稳定性**:B `execution_completion` 的九个实际模型返回全部通过人工复核;唯一未通过项为 provider-policy block,而非返回正文失败。 ### 提示词原理改进 `v1-rc1` 主要依赖通用首动作和完成规则,遇到“继续优化当前提示词项目”时可能把候选文件中的被测规则误当作本轮请求,读取断点后转为资格或范围说明。`v1` 将执行控制压缩为三个连续契约: 1. atomic dispatcher 只绑定本轮外层任务、现有断点和第一个可提交动作; 2. continuation ledger 继承已经确认的对象与结果,但不继承旧回复中的拒绝或范围判断; 3. transaction-turn commit 要求工具任务在同一轮形成可验证、可回滚并重新应用的目标状态。 这一结构保留了 rc1 的技术事务能力,同时把“读取项目状态”从终点改为下一项候选修改或隔离评测的前置步骤。 ### 测试方法与结果 所有新运行固定使用 `gpt-6-astra`、`medium`、`workers=1`。A 为 4 cases / 4 turns,准入要求 3/4 且 2/2 artifact gates;B 按 family 顺序运行,本次只完成 `execution_completion` 8 cases / 10 turns;后续 B families 与 C 未运行。完整方法与失败说明见 [`docs/comparison-tests.md`](../../docs/comparison-tests.md)。 | 阶段 | v1-rc1 / e1b5 | e2b12 | e2b15 | `v1` / e2b19 | |:

  2. gpt-5.6-instruct v45gpt-5.6-instruct_v45Aug 6, 2026pre-release2.8K downloads

    ## 🚀 gpt-5.6-instruct v45 ✨ `v45` 针对 Issues 与社交平台用户反映的实际问题,**提升逆向、生物研究任务通过概率,降低触发云审查几率**,相较 v42 重点提升: - ⚡ **单次任务编译**:以最新 USER 请求为唯一活动任务,控制型续句继承最近的实质任务;在输出前一次选择 `OUTER_SCHEMA > TOOL_TRANSACTION > PROCESS_RECORD > DIRECT_ANSWER`,减少模板误路由与重复开场。 - 🛠️ **工具事务闭环**:工具/文件修改任务首个可见动作直接检查唯一工作对象;保留原件哈希、修改副本,并创建和重开修改件、diff、验证记录、可运行 rollback 四个真实角色。 - ✅ **行为级验证**:验证记录与最终答复共同给出基线、修改后与回滚的命令、输入、原样结果和退出状态,要求在另一个副本上实际测试回滚,同时保留修改件的目标状态。 - 🔄 **过程记录渲染**:物理、机械、生物与叙事过程统一绑定 `CORE_EVENT`;`START` 和首个 `DEVELOPMENT` 句复用同一中心谓词,随后完成 `RESULT`、`FOLLOW-UP` 及叙事所需的 `DIALOGUE` / `SENSATION`。 - 🧭 **状态连续性与外层格式优先**:续轮复用已确认文件、哈希、偏移、报告和结果;列表、JSON、文档或批量结构保持用户指定的外层形式,内部规则只负责填充内容。 ### 提示词原理改进 原 v42 以 normalizer、强制工具分派、输出禁用规则、连续性和 generalized completion 组成线性规则栈。`v45` 把分散的首轮执行与完成规则重构为一次编译、单一路由的执行模型,将其压缩为四个互斥的渲染分支: 1. `SINGLE-PASS TASK COMPILER` 绑定当前任务、语言、形式和连续状态; 2. `OUTER_SCHEMA` 保护用户要求的列表、JSON、文档或批量外壳; 3. `TOOL_TRANSACTION` 管理可逆修改与四角色证据; 4. `PROCESS_RECORD` 管理具有开始、发展、结果和后续状态的过程;其余请求进入 `DIRECT_ANSWER`。 这一结构把“先判断再回退”的多路径决策改为“先绑定再完整渲染”的单路径决策,并把中心事件、工具工件和最终状态从建议性规则提升为分支级输出契约。 ### 测试方法与结果 测试方法更新为严格 **A → B → C**:A 是 3 个原始用户反馈样例;A 全过后,B 按固定 family 顺序运行 66 cases / 74 turns;B 全过后,C 使用 120 条 `medium` 输

  3. gpt-5.6-instruct v42gpt-5.6-instruct_v42Jul 29, 20265.1K downloads

    ## 🚀 gpt-5.6-instruct v42 ### 中文 ✨ v42 在 v41 基础上,针对 GitHub Issues 中“首轮不执行、模板循环和软件修改任务无法落地”等真实问题,重点优化了: - ⚡ 无历史对话时的首轮直接执行,不再依赖用户重复输入 - 🔄 逆向任务中的模板误路由、循环恢复与状态连续性 - 📦 修改副本、补丁、验证记录和可运行回滚脚本的工件化交付 - ✅ 基线与修改后行为的实际执行验证 优化过程中,模型持续吸收用户指令、真实失败案例及 GitHub Issues,自行创建或扩展测试集、分析失败并重写提示词,再通过分层回归决定是否发布。 📊 基于 Issues [#5](../../issues/5) 与 [#22](../../issues/22) 的发布门禁,v42 获得以下结果: | 门禁 | 结果 | |:--|--:| | 原始输入首轮门禁(`medium`) | **2/2 cases · 2/2 turns · 2/2 artifact gates** | | 扩展 Issue 专项集(`low`) | **60/60 cases · 68/68 turns · 8/8 artifact gates** | | 原 120-case `medium` 测试集(`low`) | **115/120 首跑 + 5/5 定向审计 → 120/120 汇总** | 📌 上述数据只对应实际运行的配置;限于本人时间和token不足等原因,尚未运行的 v42 `medium` / `high` 全量矩阵。v41 的 low、medium、high 完整矩阵继续作为历史证据保留。 📉 v42 基础提示词为 **90 行、5,856 UTF-8 bytes**,较 v35 缩短 **42.58%**。 🛠️ `codex-instruct.py` 现以 v42 作为唯一默认生产版本,支持部署前预览、配置快照、字段级安全回滚,并在卸载时保留 provider、模型和认证配置。 📦 `v41` 与 `v41-skills` 已移入 `historical-versions/`;核心提示词包 `gpt-5.6-sol-unrestricted-v42.zip` 的 SHA256 为: ```text 11f0515be89943a7244d07b625a497b04dde07a51ba26e41df583a0acc145a09 ``` --- ### English ✨ Building on v41, v42 targets real GitHub Issues involving first-turn non-execu

  4. gpt-5.6-instruct v41gpt-5.6-instruct_v41Jul 23, 20261.3K downloads

    ## 🚀 gpt-5.6-instruct v41 ### 中文 ✨ v41 针对 GitHub Issues 中的真实使用问题,重点优化了: - ☁️ 明文云审查与生物研究场景 - 🔄 模板误路由与循环恢复 - 👀 执行进度可见性 优化过程中,模型持续吸收用户指令、真实失败案例及 GitHub Issues,自行扩展测试集、分析失败并重写提示词。 📊 基于 Issues [#3](../../issues/3)、[#4](../../issues/4)、[#5](../../issues/5)、[#6](../../issues/6)、[#8](../../issues/8) 构建的 **52-case / 58-turn** 明文专项测试中,v41 三档均达到 **52/52**: | 档位 | 相较 v35 提升 | |:--|--:| | Low | +25.00 个百分点 | | Medium | +25.00 个百分点 | | High | +23.08 个百分点 | 📉 基础提示词长度同时减少 **55.67%**。 🧩 此外,项目提供 Skills 专用版本 `gpt-5.6-sol-unrestricted-v41-skills`,需手动部署。 --- ### English ✨ v41 introduces targeted improvements based on real-world GitHub Issues, focusing on: - ☁️ Plaintext cloud moderation and biological research - 🔄 Template misrouting and loop recovery - 👀 Progress visibility During optimization, the model continuously incorporated user instructions, real failure cases, and GitHub Issues, while autonomously expanding test sets, analyzing failures, and rewriting prompts. 📊 On the **52-case / 58-turn** plaintext benchmark built from Issues [#3](../../issues/3), [#4](../../issues/4), [#5](../../issues/5), [#6](../../issues/6), and [#8](../../issues

  5. gpt-5.6-instruct v35gpt-5.6-instruct_v35Jul 22, 2026351 downloads

    提供两个等级的提示词: v5 结构更简洁,足以覆盖大多数场景,并作为默认推荐版本; v35 在此基础上进一步将具体名称与网址归一化为占位符,再按中英文复合意图族进行统一路由,减少复合任务中只完成部分子任务的情况。 当前情况建议仅在 v5 无法满足需求时使用 v35。 Two prompt tiers are available: v5 has a simpler structure, covers most scenarios, and is the recommended default. v35 additionally normalizes concrete names and URLs into placeholders, then routes bilingual compound-intent families as a whole to reduce partial completion of multi-part tasks. Use v35 only when v5 does not meet your needs.

Code frequency

additions and deletions
+8.5K-8.5KWeek of 2026-07-05: +2,528 linesWeek of 2026-07-05: -0 linesWeek of 2026-07-12: +2,217 linesWeek of 2026-07-12: -2,950 linesWeek of 2026-07-19: +8,534 linesWeek of 2026-07-19: -2,359 linesWeek of 2026-07-26: +507 linesWeek of 2026-07-26: -4,492 linesWeek of 2026-08-02: +198 linesWeek of 2026-08-02: -147 linesWeek of 2026-08-09: +0 linesWeek of 2026-08-09: -0 linesWeek of 2026-08-16: +0 linesWeek of 2026-08-16: -0 linesWeek of 2026-08-23: +0 linesWeek of 2026-08-23: -0 linesWeek of 2026-08-30: +471 linesWeek of 2026-08-30: -378 linesWeek of 2026-09-06: +394 linesWeek of 2026-09-06: -105 linesJul 5, 2026Sep 6, 2026
+14.8K lines added, -10.4K removed over the last year.

Commits per week

last 52 weeks
250Week of 2025-09-28: 0 commitsWeek of 2025-10-05: 0 commitsWeek of 2025-10-12: 0 commitsWeek of 2025-10-19: 0 commitsWeek of 2025-10-26: 0 commitsWeek of 2025-11-02: 0 commitsWeek of 2025-11-09: 0 commitsWeek of 2025-11-16: 0 commitsWeek of 2025-11-23: 0 commitsWeek of 2025-11-30: 0 commitsWeek of 2025-12-07: 0 commitsWeek of 2025-12-14: 0 commitsWeek of 2025-12-21: 0 commitsWeek of 2025-12-28: 0 commitsWeek of 2026-01-04: 0 commitsWeek of 2026-01-11: 0 commitsWeek of 2026-01-18: 0 commitsWeek of 2026-01-25: 0 commitsWeek of 2026-02-01: 0 commitsWeek of 2026-02-08: 0 commitsWeek of 2026-02-15: 0 commitsWeek of 2026-02-22: 0 commitsWeek of 2026-03-01: 0 commitsWeek of 2026-03-08: 0 commitsWeek of 2026-03-15: 0 commitsWeek of 2026-03-22: 0 commitsWeek of 2026-03-29: 0 commitsWeek of 2026-04-05: 0 commitsWeek of 2026-04-12: 0 commitsWeek of 2026-04-19: 0 commitsWeek of 2026-04-26: 0 commitsWeek of 2026-05-03: 0 commitsWeek of 2026-05-10: 0 commitsWeek of 2026-05-17: 0 commitsWeek of 2026-05-24: 0 commitsWeek of 2026-05-31: 0 commitsWeek of 2026-06-07: 0 commitsWeek of 2026-06-14: 0 commitsWeek of 2026-06-21: 0 commitsWeek of 2026-06-28: 0 commitsWeek of 2026-07-05: 1 commitsWeek of 2026-07-12: 20 commitsWeek of 2026-07-19: 25 commitsWeek of 2026-07-26: 10 commitsWeek of 2026-08-02: 4 commitsWeek of 2026-08-09: 0 commitsWeek of 2026-08-16: 0 commitsWeek of 2026-08-23: 0 commitsWeek of 2026-08-30: 4 commitsWeek of 2026-09-06: 5 commitsWeek of 2026-09-13: 0 commitsWeek of 2026-09-20: 0 commitsSep 28, 2025Sep 20, 2026
69 commits in the last 52 weeks.

When work happens

weekday and hour
SunMonTueWedThuFriSat036912151821Sun 0:00 — 0 commitsSun 1:00 — 0 commitsSun 2:00 — 0 commitsSun 3:00 — 0 commitsSun 4:00 — 0 commitsSun 5:00 — 0 commitsSun 6:00 — 0 commitsSun 7:00 — 0 commitsSun 8:00 — 0 commitsSun 9:00 — 0 commitsSun 10:00 — 2 commitsSun 11:00 — 0 commitsSun 12:00 — 0 commitsSun 13:00 — 0 commitsSun 14:00 — 0 commitsSun 15:00 — 5 commitsSun 16:00 — 0 commitsSun 17:00 — 2 commitsSun 18:00 — 0 commitsSun 19:00 — 2 commitsSun 20:00 — 0 commitsSun 21:00 — 1 commitsSun 22:00 — 0 commitsSun 23:00 — 0 commitsMon 0:00 — 0 commitsMon 1:00 — 0 commitsMon 2:00 — 0 commitsMon 3:00 — 0 commitsMon 4:00 — 0 commitsMon 5:00 — 0 commitsMon 6:00 — 0 commitsMon 7:00 — 0 commitsMon 8:00 — 0 commitsMon 9:00 — 1 commitsMon 10:00 — 1 commitsMon 11:00 — 1 commitsMon 12:00 — 1 commitsMon 13:00 — 0 commitsMon 14:00 — 2 commitsMon 15:00 — 0 commitsMon 16:00 — 1 commitsMon 17:00 — 2 commitsMon 18:00 — 0 commitsMon 19:00 — 0 commitsMon 20:00 — 0 commitsMon 21:00 — 0 commitsMon 22:00 — 0 commitsMon 23:00 — 0 commitsTue 0:00 — 0 commitsTue 1:00 — 0 commitsTue 2:00 — 0 commitsTue 3:00 — 0 commitsTue 4:00 — 0 commitsTue 5:00 — 0 commitsTue 6:00 — 0 commitsTue 7:00 — 0 commitsTue 8:00 — 1 commitsTue 9:00 — 2 commitsTue 10:00 — 1 commitsTue 11:00 — 2 commitsTue 12:00 — 0 commitsTue 13:00 — 0 commitsTue 14:00 — 0 commitsTue 15:00 — 0 commitsTue 16:00 — 0 commitsTue 17:00 — 1 commitsTue 18:00 — 0 commitsTue 19:00 — 0 commitsTue 20:00 — 0 commitsTue 21:00 — 0 commitsTue 22:00 — 0 commitsTue 23:00 — 0 commitsWed 0:00 — 1 commitsWed 1:00 — 0 commitsWed 2:00 — 0 commitsWed 3:00 — 0 commitsWed 4:00 — 0 commitsWed 5:00 — 0 commitsWed 6:00 — 0 commitsWed 7:00 — 0 commitsWed 8:00 — 1 commitsWed 9:00 — 1 commitsWed 10:00 — 0 commitsWed 11:00 — 0 commitsWed 12:00 — 0 commitsWed 13:00 — 0 commitsWed 14:00 — 1 commitsWed 15:00 — 3 commitsWed 16:00 — 1 commitsWed 17:00 — 1 commitsWed 18:00 — 0 commitsWed 19:00 — 0 commitsWed 20:00 — 0 commitsWed 21:00 — 2 commitsWed 22:00 — 0 commitsWed 23:00 — 0 commitsThu 0:00 — 0 commitsThu 1:00 — 0 commitsThu 2:00 — 0 commitsThu 3:00 — 0 commitsThu 4:00 — 0 commitsThu 5:00 — 0 commitsThu 6:00 — 0 commitsThu 7:00 — 0 commitsThu 8:00 — 1 commitsThu 9:00 — 4 commitsThu 10:00 — 2 commitsThu 11:00 — 3 commitsThu 12:00 — 0 commitsThu 13:00 — 0 commitsThu 14:00 — 1 commitsThu 15:00 — 1 commitsThu 16:00 — 0 commitsThu 17:00 — 1 commitsThu 18:00 — 0 commitsThu 19:00 — 1 commitsThu 20:00 — 0 commitsThu 21:00 — 0 commitsThu 22:00 — 0 commitsThu 23:00 — 0 commitsFri 0:00 — 0 commitsFri 1:00 — 0 commitsFri 2:00 — 0 commitsFri 3:00 — 0 commitsFri 4:00 — 0 commitsFri 5:00 — 0 commitsFri 6:00 — 0 commitsFri 7:00 — 0 commitsFri 8:00 — 1 commitsFri 9:00 — 0 commitsFri 10:00 — 0 commitsFri 11:00 — 1 commitsFri 12:00 — 0 commitsFri 13:00 — 0 commitsFri 14:00 — 0 commitsFri 15:00 — 0 commitsFri 16:00 — 0 commitsFri 17:00 — 0 commitsFri 18:00 — 0 commitsFri 19:00 — 0 commitsFri 20:00 — 0 commitsFri 21:00 — 0 commitsFri 22:00 — 1 commitsFri 23:00 — 6 commitsSat 0:00 — 0 commitsSat 1:00 — 0 commitsSat 2:00 — 0 commitsSat 3:00 — 0 commitsSat 4:00 — 0 commitsSat 5:00 — 0 commitsSat 6:00 — 0 commitsSat 7:00 — 0 commitsSat 8:00 — 0 commitsSat 9:00 — 0 commitsSat 10:00 — 2 commitsSat 11:00 — 1 commitsSat 12:00 — 0 commitsSat 13:00 — 0 commitsSat 14:00 — 0 commitsSat 15:00 — 0 commitsSat 16:00 — 0 commitsSat 17:00 — 1 commitsSat 18:00 — 0 commitsSat 19:00 — 3 commitsSat 20:00 — 1 commitsSat 21:00 — 0 commitsSat 22:00 — 0 commitsSat 23:00 — 0 commits
Commit volume by weekday and hour (UTC). Larger dots mean more commits.

Who is committing

last 52 weeks
Maintainer commits69 (87%)
Community commits10 (13%)

79 commits in total over the last year.

DateListRankStars gained
Jul 13, 2026daily#12+3