SumSec-Skills

用于维护 SumSec-Skills 多平台插件元数据与发布清单：package.json、plugin.json、Claude/Cursor/Codex manifest、marketplace JSON、版本号、描述、关键词、安装文档。

管理 submodule skill 与插件同步：添加/删除映射、运行同步脚本。sync skills、sync plugins、skill-map.json、--add-plugin、从 submodule 导入

Read and send mails from a cloudflare_temp_email mailbox using a user-supplied Address JWT and API base URL. Use when the user (or an agent such as OpenClaw / Codex / Cursor) needs to list the inbox, fetch a specific message, or send an email via the server-parsed /api/parsed_mails, /api/parsed_mail/:id, and /api/send_mail endpoints. Falls back to local parsing of /api/mail/:id raw source with mail-parser-wasm + postal-mime if the parsed endpoints are unavailable. Does NOT handle mailbox creation - the user provides the JWT themselves.

SummerSec personal Agent Skills collection: writing-zh, media-tools, dev-tools, agents-dev, cloudflare-email, taste-skill.

SummerSec personal Agent Skills collection. Contains reusable SKILL.md files organized by category. Use when you need skills from writing-zh, media-tools, dev-tools, agents-dev, cloudflare-email, or taste-skill plugins.

将项目根目录的 Claude 记忆文件统一为指向根目录 AGENTS 文件的软链接。用于用户要求把 `CLAUDE.md` 或 `claude.md` 指向 `AGENTS.md` 或 `agents.md`、希望批量统一多个项目的根目录记忆文件，或需要兼容 Windows、macOS、Linux 的跨平台软链接处理流程时。

Whole-repo audit for over-engineering. Scans the codebase and ranks what to delete, simplify, or replace with stdlib/native equivalents. Use when the user asks to audit a repo for bloat, unnecessary abstractions, redundant dependencies, or "what can I delete from this codebase".

Quick reference for ponytail modes and companion skills. One-shot display, not a persistent mode. Use when the user asks how to use ponytail, which modes exist, or what ponytail commands/skills are available.

Review code or diffs only for over-engineering. Finds what to delete, replace with stdlib/native features, inline, or shrink. Use when the user asks whether something is over-engineered, what can be deleted, how to simplify a diff, or explicitly invokes ponytail review.

Forces the laziest solution that actually works: YAGNI first, reuse code that already exists, prefer stdlib and native platform features, avoid new dependencies, and stop at the smallest correct implementation. Use when the user asks for the simplest path, minimal code, fewer dependencies, less boilerplate, less over-engineering, or explicitly mentions ponytail, lazy mode, YAGNI, shortest path, or "do less".

为 SumSec / sumsec.me 中文技术文章设计或生成正文配图、shot list、插图提示词、图片提示词提取和已有图修改建议。用于安全研究、漏洞复盘、AI Agent、Skill 仓库、工程工作流、方法论、流程、结构、状态、隐喻或观点的“手绘配图”“文章插图”“参考图提取人物特征”“提取这张图的提示词”“个人形象 prompt”“角色设定图 prompt”“SumSec Observer 角色”“稳定人物色彩”“胸前 SummerSec 铭牌”“去标题/改图”等任务；默认使用 SumSec Observer 个人化原创角色：年轻成人安全研究员 / 系统观测员，纯白手绘、深炭干净轮廓线、低饱和人物局部上色、深墨凌乱短发、细框眼镜、浅冷灰高领连帽夹克、暗青蓝内衬与斜挎包带、灰褐工具包、黑色夹板、日志纸、胸前小工作证式 SummerSec 铭牌、双 S 戒指、清醒低情绪的工作中忧郁感；必须干净下颌、无胡子/胡茬，不走阳光笑脸、老成大叔、吉祥物、可爱小人、黑客反派或外部 IP 路线。

Use when creating, reviewing, or verifying Claude Agent SDK applications in Python or TypeScript, including new SDK app scaffolding guidance, SDK setup checks, verifier-agent style reviews, dependency/version checks, and best-practice audits.

优化、审查或诊断 Agent Skills（SKILL.md）。用于 skill 质量分析、重构技能、skills 不触发、skill 没生效、optimize skills、analyze skills、check my skills。

当用户要求创建流程编排类 Skill、管线 Skill、工作流 Skill、多步骤分析 Skill，或提及流程编排、pipeline skill、workflow skill、步骤框架、进度文件、任务编排、审计流程、分析管线、复杂任务 skill时，应使用此技能。提取自 biz-vul-security 的六大架构模式（进度文件驱动、步骤框架分离、三阶段执行、脚本自动化、资源分层、路径一致性），提供从需求梳理到交付自检的完整方法论。

按 YYYY-MM-DD 在 Windows/macOS/Linux 检索本机聊天与 Agent 会话历史（Claude Code 项目 jsonl、Codex CLI history.jsonl、Cursor workspaceStorage 与 state.vscdb）；可只导出用户提示词。用户要查看或导出某日记录、按天筛 jsonl/SQLite、只抽提示词、排查历史丢失或做只读审计时使用。

Use when the user wants to generate, create, or draw images using AI; mentions text-to-image, AI drawing, image generation, 画图, 生成图片, AI 绘画, or needs visual content from text descriptions. Supports any OpenAI-compatible image API (OpenAI, Azure, Right.Codes, etc.). Requires OPENAI_API_KEY or GPT_API_TOKEN environment variable.

用于把本仓库 Markdown 文章转成网页版 PPT、HTML slides、演讲稿页面或独立演示页：同目录、同 basename、SUMSEC 回链、全屏按钮、画布不超视口、slide-writer 主题流程。

去 AI 味、降 AIGC、人性化、humanize chinese、论文降重、知网维普万方、AI 检测、文本去机器味。 适用于中文文本改写、AI 痕迹检测、学术降重、风格转换；有 Python 时优先用本 skill 内零依赖 CLI，无 CLI 时按深度指南改稿。

用于安全提交、推送或创建 PR/MR：提交代码、推送分支、创建 PR、创建 MR、提交PR、提交MR、发MR、commit and PR、commit and push。

用于编写、审查或排错 Remotion（React 程序化视频）代码：composition、时间轴、interpolate、字幕、资源、图表、转场、FFmpeg、静音检测、create-video、Remotion Studio。