secure-engineering-skill-pack - An AI skill for building modular, maintainable sof

项目说明

Secure Engineering Skill Pack

简体中文

A focused AI skill pack for e implementation, secure design, dependency and privacy review, operational readiness, safe migrations, release decisions, and authorized defensi review.

Included Skills

modular-e

Use for feature implementation, bug fixes, refactoring, architecture boundaries, repository e, focused tests, and o mixed-responsibility files.

release-readiness-e

Use before staging or production release, deployment approval, migration or feature-flag rollout, capability-boundary , rollback, recovery, and Go or No-Go decisions.

authorized-security-review

Use for defensi review of authentication, recovery, authorization, administrator actions, tenant isolation, abuse controls, sensitive data, and scoped security testing.

threat-modeling-lite

Use before implementing security-sensiti to identify assets, entry points, trust boundaries, abuse cases, controls, tests, and residual risk.

supply-chain-and-dependency-review

Use when adding or upgrading packages, SDKs, CI actions, build scripts, Docker images, registries, signing, and release-pipeline components.

data-pri

Use when collecting, storing, logging, exporting, deleting, analyzing, uploading, or sharing user and sensitive data.

obseradiness

Use when production workflows need logs, metrics, traces, alerts, dashboards, runbooks, containment, rollback, recovery, and data repair.

migration-and-rollback-review

Use for database, schema, data-format, queue, cache, configuration, storage, protocol, and mixed- rollout changes.

Why Focused Skills?

Each skill has a narrow trigger and loads detailed references/ only when needed. This reduces o, keeps agent context focused, and allows each skill to be installed or used independently.

Installation

Copy each desired directory from skills/ into the target tool's skill directory.

Codex

User-level:

Repository-level:

Example:

Claude Code

Personal:

Project:

Cursor, Trae, Windsurf, and GitHub Copilot

Add the relevant SKILL.md as a project rule, workspace instruction, agent rule, or custom instruction. Keep its references/ directory a when the tool supports referenced files.

Installation

Run the repository :

Then try the task in examples/quickstart.md. Confirm the agent selects the relevant focused skills, distinguishes e from assumptions, reports actually run, and does not claim unsupported release readiness.

GitHub Actions runs the same for pushes and pull requests.

Repository Structure

Every skill directory contains a short SKILL.md, optional agents/openai.yaml, and detailed guidance in references/.

Shared Resources

Core Principles

The 800-line limit is a deliberately generous review trigger, not a claim that every file over 800 lines is inherently wrong. Detailed rationale, exceptions, and team-le guidance live in the modular implementation reference.

Examples

Start with examples/quickstart.md for a complete before-and-after example.

License

MIT

项目信息

开发语言：Python

Star 数：1

技术要点

该项目属于防护思路方向的开源资源，代码结构清晰，适合安全研究人员学习参考。项目围绕An AI skill for building modular, maintainable software with evidence-driven development, secure release validation, and defensive security review.展开，提供了可运行的实现，对理解相关安全机制、工具原理与攻防技术具有较高的研究价值。

从技术实现角度看，此类项目的核心在于对底层原理的深入理解与工程化落地。开发者在研究过程中可以结合官方文档与社区讨论，梳理项目的初始化流程、核心模块职责与数据流转，逐步掌握其中的技术细节与设计思路。

使用与学习建议

1. 下载源码压缩包，解压到本地环境。

2. 根据项目类型准备对应的运行环境与依赖。

3. 按照 README 中的步骤完成配置与部署。

4. 在隔离的测试环境中运行验证，避免影响生产系统。

免责声明

本资源来源于开源社区，仅供安全研究、技术学习与合法授权范围内的攻防演练使用，请勿用于任何非法或侵害他人权益的用途。使用本项目产生的一切后果由使用者自行承担。