Grill Pro：在写代码前用追问把设计与术语钉死
Grill Pro（@user_dc515faf/grill-stan）是 SkillHub 社区 ai-agent 类 Skill，不写实现代码，而是通过「grill me」或描述计划触发串行追问，必要时自行查阅仓库，将术语写入 CONTEXT.md，并在不可逆决策时提示 ADR。本文基于 SkillHub 官方简介与安装文档，说明定位、核心能力、skillhub install 安装路径及典型使用步骤，适合在开工前压力测试方案与统一领域语言。
Read Moregrill-with-docs：用追问把方案磨清，并同步沉淀 CONTEXT 与 ADR
grill-with-docs 是 SkillHub 社区 Skill（@user_3c6cb52e/grill-with-docs，v1.0.0，ai-agent，无需 API Key）。它要求 Agent 对计划或设计逐轮单问深挖，能查代码则先查代码，并在会话中维护 CONTEXT.md 术语表与 docs/adr/ 决策记录。包内含 CONTEXT/ADR 格式说明、参考文档及三个 Python 标准库校验脚本。本文介绍定位、核心纪律、skillhub 安装到 Cursor/Claude/Codex skills 目录的方法，以及典型启动方式与使用注意。
Read MoreDSH Plugin: dsh-domain-modeling (Domain Modeling and Knowledge Maintenance)
dsh-domain-modeling is a DeepSeek Harness plugin designed to address context management challenges in agent development. Maintained by satan9394, it offers core features such as domain modeling, maintaining project glossaries, and recording Architecture Decision Records (ADRs). Inspired by mattpocock/skills, it helps agents generate or update CONTEXT.md files, align terminology understanding, and author ADR documentation to trace decision context. It is well suited to long-term maintenance projects with complex knowledge systems and effectively reduces the risk of context loss.
Read More# 翻译结果： Architecture Decision Records: Documenting Technical Decisions with Agent Skill
### Translation: Let me introduce the *architecture-decision-records* skill included in awesome-cursor-skills: it teaches AI Agents to document technical decisions in accordance with ADR specifications, covering their backgrounds, alternative options and justifications. The skill covers when to write ADRs, template structures, a six-step workflow, installation methods and typical prompt examples, helping teams turn architectural governance from verbal agreements into retrievable decision logs.
Read More