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 instructions/timzaak/web-dev-skills/agents-mdgit clone --depth 1 https://github.com/timzaak/web-dev-skillsWhat 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 | $0.01191 | $0.01191 |
| Opus 5 | $0.00596 | $0.00596 |
| Sonnet 5 | $0.00238 | $0.00238 |
| Haiku 4.5 | $0.00119 | $0.00119 |
Grade A, and why
web-dev-skills AGENTS.md 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 3d 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 — 54 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Skill 编写规范
本仓库用于开发 Claude Code plugin t-tools。在这里修改的是可触发、可执行、可恢复、可验收的 AI 工程工作流,不是目标项目的业务代码。
规则归属
skills/负责阶段编排、前置校验、状态推进、产物写入和失败恢复。agents/负责单一角色的具体执行或只读验收,不承担跨阶段编排。protocols/是跨 skill 和 agent 的共享契约源;状态结构、输出结构、评分规则和调度规则只在这里定义。guides/承载工程规范和领域实践;skill 只说明何时读取,不复制大段内容。scripts/承载需要重复执行或必须保持确定性的操作。- 同一规则只维护一份。先确定规则归属,再通过语义路径引用它。
编写 Skill
- 一个 skill 只负责一个边界清晰的阶段或能力。不要把相邻阶段的实现、测试和验收揉进同一个入口。
name使用稳定的 kebab-case;description同时说明“做什么”和“何时使用”,并与容易混淆的相邻 skill 明确消歧。- 只写会改变 AI 决策或执行结果的内容。删除通用知识、仓库文件系统已经显而易见的信息和没有实际约束力的说明。
- 根据风险控制自由度:开放性工作给出目标和判断标准;脆弱、可破坏或必须一致的操作才规定固定顺序、命令或脚本。
SKILL.md保留入口所需的目标、适用范围、核心流程、门禁、失败处理和资源路由。详细契约放入protocols/,工程知识放入guides/,可执行逻辑放入scripts/。- 引用必须说明何时读取或执行。避免深层引用链,也不要默认加载当前任务不需要的所有参考文件。
- 涉及阶段产物时,明确输入来源、输出位置、成功条件和下游消费条件。
- 涉及持久化状态时,明确状态转换、可恢复点、重复执行行为以及失败后由谁继续。
- 需要用户裁决的缺口必须暴露给用户;不要把未确认假设写入产物后继续推进。
- 面向目标项目引用插件内部文件时使用
${CLAUDE_PLUGIN_ROOT},不要依赖插件安装位置或当前工作目录。
编排与角色边界
- Skill 负责决定下一步、选择 agent、传递必要上下文并汇总结果;不要在 skill 和 agent 中重复实现同一判断逻辑。
- Dev agent 可以实现;test agent 专注测试;accept agent 默认只读验收并输出可复核证据。
- Agent 输入、输出或评分被多个入口使用时,抽取为
protocols/契约,不在各调用方分别描述。 t-prd-check、t-design-check、t-task-check是可选质量检查。可以在高风险或 AI 输出不稳定时推荐,但不得悄悄变成主链路的强制前置条件。- Accept 是实现后的验收收口。增加前置检查不能替代或弱化 accept 的职责。
- 纯技术且不改变业务逻辑、产品规则、用户流程或验收目标的方案,才可以绕过 PRD 直接从技术预研进入设计。
修改前后检查
- 修改 skill 前先搜索它引用的 protocol、guide、agent、script,以及引用该 skill 或其产物的下游入口。
- 修改共享状态、输出、评分或调度规则时,更新对应
protocols/,并检查所有消费者;不要只修当前调用方。 - 修改目标项目运行时文件结构时,先检查
${CLAUDE_PLUGIN_ROOT}/protocols/runtime-boundaries.md和${CLAUDE_PLUGIN_ROOT}/skills/t-init/的相关模板。 - 修改命令调用、skill 名称、公开工作流或安装方式时,同步更新
README.md和README.en.md。 - 修改插件暴露能力或依赖时,同步检查
.claude-plugin/plugin.json。 - 中文文档是主说明;存在对应英文文档时保持语义一致。
验证
- 把 skill 文本视为行为代码验证,不以 Markdown 可读性代替行为正确性。
- 按变更风险验证正常路径、前置条件缺失、下游门禁、失败恢复和重复执行;只验证受影响的场景,不编造无关测试。
- 新增或修改脚本时直接运行脚本,并覆盖本次变更涉及的错误路径。
- Markdown 链接使用
${CLAUDE_PLUGIN_ROOT}/scripts/check-markdown-links.py检查。 packages/playwright-unified-logger/的改动在该 package 目录内运行对应 npm 脚本。- 发布仍由
${CLAUDE_PLUGIN_ROOT}/scripts/release.py和/t-tools:t-release约束:语义版本不带v,git tag 使用v前缀。 - 仓库可能有用户未提交改动;不要回滚或覆盖与当前任务无关的内容。
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.
- 3d ago First seen · 54 lines · 1,191 tokens per session scan A 208c096e2b80
web-dev-skills AGENTS.md is an instructions file published in the GitHub repository timzaak/web-dev-skills (69 stars, last pushed 4d ago), licensed Apache-2.0. It adds 1,191 tokens to every session, about $0.0060 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 2026-08-30.
Other instructions, from other repositories
superpowers-zh CLAUDE.md
Instructions for jnMetaCode/superpowers-zh, covering superpowers — 贡献者指南, 如果你是 ai agent, pull request 要求, 我们不会接受的内容 and 第三方依赖.
superpowers-zh GEMINI.md
Instructions for jnMetaCode/superpowers-zh, covering superpowers-zh 中文增强版, 核心规则, 可用 skills and 如何使用.
spec-superflow GEMINI.md
Instructions for MageByte-Zero/spec-superflow, covering spec-superflow and spec-superflow v1.0.1 | 阶段: {{state}} | 工作流: {{workflow}}.
spec-superflow copilot-instructions.md
Instructions for MageByte-Zero/spec-superflow, covering phase guard, ✅ 允许操作, ⛔ 禁止操作 and 🔔 决策点.
mlops-coding-skills AGENTS.md
Instructions for MLOps-Courses/mlops-coding-skills, covering agents.md, project overview, setup & core commands, definition of done and conventions & idioms.
claude-skills CLAUDE.md
Instructions for mastepanoski/claude-skills, covering claude.md, repository purpose, testing skills locally, repository architecture and core structure.