Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/yangfan-code-star/context-dev/initnpx skills add YangFan-Code-Star/context-dev --skill initgit clone --depth 1 https://github.com/YangFan-Code-Star/context-devWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00112 | $0.05066 |
| Opus 5 | $0.00056 | $0.02533 |
| Sonnet 5 | $0.00022 | $0.01013 |
| Haiku 4.5 | $0.00011 | $0.00507 |
Grade A, and why
init scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 199 lines — stays where its author put it; the contents beside it link to each section on GitHub.
初始化项目
模板本身是空的。这个技能的任务是通过提问把它变成一份准确的项目手册——问完之后,一个半年后第一次进来的 agent 应该只读 AGENTS.md 就能正确干活。
产出质量取决于问题质量。宁可多问两轮,也不要靠猜填占位符。
访谈原则
这几条决定了访谈是好用还是折磨人:
- 一次问 3~5 个,不要一次抛 20 个。 问完一组、得到回答、落盘,再问下一组。
- 每个问题都给推荐答案。 让用户能直接说"就用默认"。开放式问题("你想怎么设计?")会让用户卡住,带选项的问题("A 还是 B,我建议 A 因为……")不会。
- 不问能自己看出来的。 见阶段 0。问用户
package.json里写着的东西会立刻消耗掉他的耐心。 - 用户说"不知道 / 你决定"时,不要卡住。 给出你的选择 + 一句理由,标注为「agent 代填,待确认」,继续往下走。
- 答案模糊就追问一次。 "要好用"、"性能要好"这种答案不可执行,追问:"能给个具体场景吗?比如什么情况下你会觉得它不好用?"
- 每阶段结束立刻落盘。 用户可能随时中断,写进文件的才算数。
- 只记录模型猜不到的东西。 用户的口径选择、历史包袱、被否决的方案、项目特有的陷阱。不要把"要写测试"、"注意错误处理"这类模型本来就知道的内容写进产出。
- 用推导式提问,少用清单式。 能从上一个答案推出来的问题,就说明推导过程再问——"你说这是单用户本地工具,那我推断没有登录鉴权;下一个问题:数据存哪?因为这决定本地文件还是嵌入式数据库。"清单式提问(把题库念一遍)让用户觉得被审问,推导式让他觉得被理解,也让误解在早期暴露。
流程
- [ ] 阶段 0:生成骨架 + 勘察现状(不问用户)
- [ ] 阶段 1:项目定位(6 个必答)
- [ ] 阶段 2:领域深挖(按形态分支)
- [ ] 阶段 3:红线与口径
- [ ] 阶段 4:范围边界
- [ ] 阶段 5:落盘、体检、复述确认
每个阶段结束跑一次收敛自检——只要下面四条都能回答,就不必按部就班走完所有阶段,直接进阶段 5 落盘:
- 能一句话说出项目做什么、给谁用
- 能说出可验证的成功标准
- 能说出第一个任务的完整执行上下文(入口、数据、红线)
- 阶段 3 的红线与口径六问都已问过(用户答"没有"也算问过;没问过就不可能"没有未决问题")
四条不全通过,就继续下一阶段。问题数量由内容决定,不由流程决定——问够了就停,没问够就继续。
轻量档:先问 3 个就开工
完整流程是给"要长期维护、别人也会用"的项目准备的。周末小工具、一次性脚本、还不确定做不做得成的试验,第一次接触就被问三十道题只会让人放弃。这种情况先走轻量档,在阶段 0 之后直接问这 3 个:
- 一句话说清做什么、谁用(阶段 1 问题 1)
- 做到什么程度算成功(阶段 1 问题 2)
- 从零到能跑起来、以及验证改动对不对,分别跑什么命令(阶段 1 问题 5)
轻量档只保留和具体功能实现相关的问题,不额外问敏感数据与红线,也不建议换完整版。
落盘时和完整流程一样删掉 AGENTS.md 顶部的「这个仓库还没初始化」引用块,但保留同一个 TODO(init) 欠账标记:没问到的占位符改写成 TODO(init): 轻量初始化未覆盖——/ship-change 首次碰到时现场追问再补。体检脚本会把 TODO(init) 报为 warning(不是"未初始化"的硬错误),其中轻量欠账单独计数,和「完整初始化没做完」区分开:完整初始化的完成线仍是 TODO(init) 归零,轻量档则是有意保留欠账。之后每次 /ship-change 碰到与本次改动相关的 TODO(init),当场问一句补上;与本次改动无关的欠账不要顺手全补。
阶段 0:生成骨架 + 勘察现状
先生成骨架,再看现状。 本技能自包含一套项目骨架模板(templates/,与 SKILL.md 同目录)。先确认 Node 可用(脚手架和体检脚本都是零依赖 Node 脚本):
node --version
先确认技能挂载,再生成骨架。 scaffold.mjs 只生成 12 个骨架文件,不安装技能;项目根没有 .agents/skills/ 时,初始化完成后 /ship-change、/maintain-context 不会被宿主发现。先用 dry-run 拿到项目根并预览:
What ships with it
15 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- references/output-map.md 5.4 KB
- references/question-bank.md 7.6 KB
- scaffold.mjs 9.5 KB runs code
- templates/.agents/evals/behavior-cases.md 4.2 KB
- templates/.gitignore 517 B
- templates/AGENTS.md.tmpl 5.9 KB
- templates/docs/architecture.md 1.6 KB
- templates/docs/decisions/0001-record-architecture-decisions.md 2.2 KB
- templates/docs/decisions/template.md 1.4 KB
- templates/docs/glossary.md 1.4 KB
- templates/docs/goals.md 2.1 KB
- templates/docs/learning-inbox.md 1.7 KB
- templates/docs/roadmap.md 1.1 KB
- templates/docs/troubleshooting.md 989 B
- templates/scripts/audit-context.mjs 20 KB runs code
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- yesterday First seen · 199 lines · 112 tokens per session scan A 27f94f2abefb
init is a skill published in the GitHub repository YangFan-Code-Star/context-dev (2 stars, last pushed 14d ago), licensed MIT. It adds 112 tokens to every session and 5,066 once invoked, about $0.0006 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
dsh-plugin-dev
开发 DeepSeek Harness (DSH) 插件的标准与权威参考:编写/修改/审查/调试 DSH/Cordis 插件、服务、事件、插件配置、模型工具、LLM 适配器、三种角色拆分、打包安装、workspace 包、cordis.yml 组合时使用;提到 DSH 插件、Cordis、plugin、服务、事件、工具、适配器即触发。 The authoritative standard for developing DeepSeek Harness (DSH) plugins — create, modify, review or debug DSH/Cordis plugins, services, events…
agent-handoff
Cross-platform skill for Codex, Claude Code, and DeepSeek Harness (DSH) that creates, updates, compacts, rotates, repairs, and reviews durable repository handoffs. Use when bootstrapping cross-session memory; creating or maintaining AGENTHANDOFF.md and .agent-handoff files; adding AGENTS.md or .claude/CLAUDE.md rules…
acontext-installer
Install Acontext, Login & Init Acontext Project, Add Skill Memory to Agent.
storage-analyzer
对 macOS 做一次只读存储分析,产出交互式 HTML 报告。流程:扫描 → 分析分级 → 生成网页 → 打开。.
brainstorming
You MUST use this before any creative work - creating features, building components, adding functionality, or modifying behavior. Explores user intent, requirements and design before implementation.
writing-plans
Use when you have a spec or requirements for a multi-step task, before touching code.