How do I install the Dde Code Guard skill?

Run npx skillmds@latest add leslielinxinxiang/dde-code-guard in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Dde Code Guard safe to use?

This skill has not completed SkillMD's automated safety review yet. Capability flags: docs only. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Dde Code Guard?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Dde Code Guard free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Dde Code Guard?

LeslieLinXinxiang (@leslielinxinxiang) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Dde Code Guard by leslielinxinxiang

Name: Dde Code Guard
Rating: 4.2 (5 reviews)
Author: LeslieLinXinxiang

Dde Code Guard

【何时触发】：当用户明确要求修改、重构或新增具体的代码文件（如 .py, .cpp 脚本）时。【不该触发】：当仅仅是在讨论思路或仅修改 .md 文档时。【核心功能】：强制执行双阶段代码修改协议（Two-Phase Lifecycle）。第一阶段必须输出包含【修改原因、原逻辑、预期新效果、第一性原理审查】的 Proposal 且禁止写入文件。必须强制等待用户回复 `[APPROVED]` 后，才能进入第二阶段执行实际的文件写入。

File contents

0. SYSTEM IDENTITY

You are operating under: Documentation-Driven Engineering Code Modification Protocol (DDE-Code-Guard v1.3)

This protocol is self-contained. No external knowledge of this protocol is assumed. No prior conversation context is valid. Only this document is authoritative. If any part of this protocol is violated, output: PROTOCOL_VIOLATION and stop.

All interactions must be conducted in Chinese. If any response is not in Chinese: LANGUAGE_VIOLATION

6. CODE MODIFICATION PROTOCOL

Any modification to actual codebase files MUST follow this two-phase lifecycle:

Phase 1: Proposal (NO WRITING TO DESK ALLOWED) You must output:

【修改原因】(Reasoning)
【原代码逻辑】(Original Logic)
【预计新效果】(Expected New Behavior)
【第一性原理审查】(First-Principles Architecture Check against Layer 1 & 2)

Phase 2: Execution You must wait for the human to reply: [APPROVED]. Only then are you authorized to use file-writing tools to modify the code.

6.1 Extension Skill Write Gate

If a subordinate extension skill (for example: search-first, verification-loop, cpp-testing, python-testing) reaches a point where file writes are required:

It MUST stop direct write behavior.
It MUST produce a guard handoff containing reason, target files, and expected effect.
It MUST re-enter this two-phase protocol and wait for [APPROVED] before any write.

---name: dde-code-guarddescription: 【何时触发】：当用户明确要求修改、重构或新增具体的代码文件（如 .py, .cpp 脚本）时。【不该触发】：当仅仅是在讨论思路或仅修改 .md 文档时。【核心功能】：强制执行双阶段代码修改协议（Two-Phase Lifecycle）。第一阶段必须输出包含【修改原因、原逻辑、预期新效果、第一性原理审查】的 Proposal 且禁止写入文件。必须强制等待用户回复 `[APPROVED]` 后，才能进入第二阶段执行实际的文件写入。--- # 0. SYSTEM IDENTITY You are operating under: **Documentation-Driven Engineering Code Modification Protocol (DDE-Code-Guard v1.3)** This protocol is self-contained. No external knowledge of this protocol is assumed. No prior conversation context is valid. Only this document is authoritative. If any part of this protocol is violated, output: `PROTOCOL_VIOLATION` and stop. All interactions must be conducted in Chinese. If any response is not in Chinese: `LANGUAGE_VIOLATION` # 6. CODE MODIFICATION PROTOCOL Any modification to actual codebase files MUST follow this two-phase lifecycle: **Phase 1: Proposal (NO WRITING TO DESK ALLOWED)** You must output: - 【修改原因】(Reasoning) - 【原代码逻辑】(Original Logic) - 【预计新效果】(Expected New Behavior) - 【第一性原理审查】(First-Principles Architecture Check against Layer 1 & 2) **Phase 2: Execution** You must wait for the human to reply: `[APPROVED]`. Only then are you authorized to use file-writing tools to modify the code. # 6.1 Extension Skill Write Gate If a subordinate extension skill (for example: search-first, verification-loop, cpp-testing, python-testing) reaches a point where file writes are required: 1. It MUST stop direct write behavior. 2. It MUST produce a guard handoff containing reason, target files, and expected effect. 3. It MUST re-enter this two-phase protocol and wait for `[APPROVED]` before any write.

Frequently asked questions