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 skills add TashanGKD/tashan-cursor-skills --skill skill-rulegit clone --depth 1 https://github.com/TashanGKD/tashan-cursor-skillsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/tashangkd/tashan-cursor-skills/skill-rule)<a href="https://agentmods.dev/skills/tashangkd/tashan-cursor-skills/skill-rule"><img src="https://agentmods.dev/badge/skills/tashangkd/tashan-cursor-skills/skill-rule/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/tashangkd/tashan-cursor-skills/skill-rule"><img src="https://agentmods.dev/badge/skills/tashangkd/tashan-cursor-skills/skill-rule.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What 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 5.1 | $0.00073 | $0.06036 |
| Opus 5 | $0.00036 | $0.03018 |
| Sonnet 5 | $0.00015 | $0.01207 |
| Haiku 4.5 | $0.00007 | $0.00604 |
Grade A, and why
skill-rule-修改规范 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 9d ago.
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 — 456 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Skill / Rule / Agent 修改规范
Skill、Rule、Agent 是 AI 行为的「操作系统」。修改它们比修改功能代码风险更高——一行错误的规则会影响所有未来任务的执行,而且问题难以溯源。
修改 Skill/Rule/Agent 和修改生产代码一样,需要相同严格的流程。
三型作用域:
.cursor/skills/:Skill 文件(步骤化流程,按场景触发).cursor/rules/:Rule 文件(全局约束,alwaysApply 或条件触发).cursor/agents/:Agent 文件(独立上下文执行单元,如 verifier / user-simulator)三型文件均受本规范约束,无例外。判断应建哪型,参照
Skill体系设计原则_v1.0.md §2.5 统一决策树。
知识导航表(执行前必须理解的概念根)
本 Skill 的步骤设计基于以下文档的核心概念。若不理解这些概念,步骤无法正确执行。
| 层级 | 文档 | 需要理解的概念 |
|---|---|---|
| D0 认知根(必读) | _内部总控/认知结构/L1_系统性文档/系统架构思维维度/Skill体系设计原则_v1.0.md |
§2.2 Rule本质(普遍约束<300词);§2.3 Skill本质(任务流程);§2.4 Agent本质(隔离执行);§2.5 三型统一决策树;§4 文档体系与知识层次 |
| D3 规范参考 | _内部总控/认知结构/L1_系统性文档/系统架构思维维度/系统对象全量分类表_20260320.md |
类型A(Rule)/ B(Skill)/ C(Agent)的完整定义与存放位置 |
| D4 运行时数据 | 目标 Skill/Rule/Agent 文件(执行时确定) | 待修改对象的当前版本和现有内容 |
核心概念速查: ① ceremony(B) = 备份→三问→最小化修改→变更记录→更新索引 ② alwaysApply Rule = 每次对话注入,违反即错误(应≤4个,目前过多是已知技术债) ③ Skill趋薄化 = 知识迁移到 K-objects,Skill 只保留导航+触发+守门
Step 0:复杂度级别判断(所有操作的第一步)
在回答三问之前,先判断本次操作属于哪个级别:
Level 1 补丁型:修改已有组件的局部内容(一个步骤/注意事项/触发词措辞)
→ 走本规范的「三问 + 备份 + 修改 + 变更记录 + 索引」流程
→ 不需要 skill-designer 引导,直接执行
Level 2 新增型:新建独立 Skill/Agent/Rule,与现有系统无显著交互
→ 先走 skill-designer 的 Step 1-5(含关卡A + 关卡C)
→ 完成后回到本规范 Step 4-6(备份 + 写入 + 变更记录 + 索引)
Level 3 集成型:新建组件,与多个现有组件有触发词重叠或调用关系
→ 先走 skill-designer 的 Step 1-7(含关卡A + 关卡B + 关卡C)
→ 完成后回到本规范 Step 4-6
Level 4 系统型:重构多个现有组件的交互关系
→ 先走 skill-designer 完整八步
→ 需要在 Step 3(产品定义)完成后,暂停等待用户二次确认规模
Level 判断有疑问时:加载 skill-designer Skill,由它引导判断。
⚠️ 修改前必须回答的三个问题(Level 1 及所有级别通用)
在动笔之前,必须能清楚回答以下三问。答不出来,不允许开始修改。
1. 这次修改的根因是什么?
→ 是 AI 执行中发现了真实的执行漏洞?
→ 还是用户/AI 凭感觉觉得"应该加点什么"?
→ 根因必须来自真实事件(某次任务执行的具体问题),不能是假设
2. 修改后,会影响哪些已有的执行路径?
→ 哪些任务会因为这个改动而改变行为?
→ 有没有可能让之前正常工作的流程变得不正常?
3. 怎么验证改完之后规则确实生效且没有副作用?
→ 能不能构造一个具体的场景来验证新规则?
→ 用什么方法检查没有破坏已有约束?
激活后立即执行
Step 1 Read: 被修改的 Skill/Rule 文件(了解现状)
Step 1.5 Read: _内部总控/认知结构/L1_系统性文档/系统架构思维维度/Skill体系设计原则_v1.0.md
→ 重点阅读:§4.3.5「认知根原则」,确认本次修改有认知根支撑
→ 带着以下问题进入 Step 2:「本次修改是否与该 Skill 的认知根一致?
是否会让这个 Skill 更难追溯到认知结构?」
Step 2 Read: 相关联的 Skill/Rule(检查是否有冲突风险)
→ 如果改测试工程师 Skill,要读 role-menu.mdc
→ 如果改 role-menu.mdc,要读所有被引用的 Skill
Step 3 回答「修改前三问」,写在修改记录里
Step 4 执行修改(见操作规范)
Step 5 执行闭环验证(见验证规范)
Step 6 写修改记录(见留痕规范)
What ships with it
8 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.
- history/SKILL_v1.0_20260319.md 6.2 KB
- history/SKILL_v1.1_20260319.md 7.7 KB
- history/SKILL_v1.2_20260319f.md 9.1 KB
- history/SKILL_v1.3_20260319sandboxfix.md 11 KB
- history/SKILL_v1.4_20260320.md 13 KB
- history/SKILL_v1.5_20260322.md 17 KB
- history/SKILL_v1.6_20260322_before_agent-scope.md 18 KB
- history/SKILL_v1.6_20260323_before_nav.md 19 KB
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.
- 9d ago First seen · 456 lines · 73 tokens per session scan A 31d3c8ad9783
skill-rule-修改规范 is a skill published in the GitHub repository TashanGKD/tashan-cursor-skills (20 stars, last pushed 5mo ago), licensed MIT. It adds 73 tokens to every session and 6,036 once invoked, about $0.0004 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 2026-09-03.
Other skills, from other repositories
retro
Engineering manager retrospective mode. Analyzes commit history, shipping velocity, and team contributions. Use when user says /retro, wants a weekly retrospective, needs to review what was accomplished, or wants team contribution analysis.
ship
Release engineer execution mode. Handles git operations, runs tests, creates PRs. Use when user says /ship, wants to commit and push, needs to create a PR, or is ready to land their changes.
github-repo-triage
Triage and maintain GitHub repositories — assess open issues/PRs, cherry-pick contributions, manage releases, and publish packages. Use when the user wants to catch up on a repo, review open issues, merge community PRs, cut a release, or publish to package registries.
gh-cli
Interact with GitHub via the gh CLI for PRs, issues, releases, Actions, and API calls. Use when performing any GitHub operation such as creating PRs, viewing issues, checking CI status, making API requests, or managing releases.
android-feature
Workflow for implementing a new Android feature, screen, fragment, activity, dialog, adapter, or enhancement in this XML/MVVM template. Use whenever the user asks to add, build, implement, or extend functionality — before writing any code. Covers planning, the MVVM/MVI skeleton, the Android configuration checklist…
template-helpers
Catalog of the reusable building blocks shipped with this Android template — BaseActivity/BaseFragment, the helpers/ extension files (navigation, lifecycle, toast, snackbar, dialogs, images, permissions, theme, locale, settings intents, date, delay), common/ (Firebase, network, observers), and…