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/ktkid/x-dev-pipeline/claude-mdgit clone --depth 1 https://github.com/KtKID/x-dev-pipelineWrote 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/instructions/ktkid/x-dev-pipeline/claude-md)<a href="https://agentmods.dev/instructions/ktkid/x-dev-pipeline/claude-md"><img src="https://agentmods.dev/badge/instructions/ktkid/x-dev-pipeline/claude-md.svg" alt="Measured on agentmods" 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 | $0.02657 | $0.02657 |
| Opus 5 | $0.01328 | $0.01328 |
| Sonnet 5 | $0.00531 | $0.00531 |
| Haiku 4.5 | $0.00266 | $0.00266 |
Grade A, and why
x-dev-pipeline CLAUDE.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 5d 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 — 107 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md
This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
仓库性质
这不是应用代码仓库，而是一个 AI 开发工作流 skill 集合——为 Claude Code / Codex 提供 /x-* 系列 slash command 的 plugin。仓库本身不包含可运行代码、没有 build、没有测试套件，所有"产物"都是 markdown skill 文件。
⚠️ 命名陷阱：仓库叫 x-dev-pipeline，仓库内部有一个子目录叫 dev-pipeline/（无 x- 前缀），它是 task 工件的输出目录。两者不是同一个东西。
仓库布局
x-dev-pipeline/
├── skills/<name>/SKILL.md # skill 定义（YAML frontmatter + 正文）
│ └── references/, templates/ # reviewer prompt、报告模板
├── dev-pipeline/tasks/<task>/ # 用 skill 时产出的 task 工件
│ ├── README.md, changelog.md, dev-report.md # qdev 轻量任务
│ ├── dev-checklist.md, diagram.md # 完整任务按需存在
│ └── reports/ # gate 报告（全部在 task 目录下）
│ ├── .fix-counter # 批量修轮数（verify 创建，fix 按轮递增，qa-gate 重置）
│ ├── verify/*.md, qa-gate/*.md # Gate ①② 报告
│ ├── fix/*.md # 修复报告（按触发节点分类）
│ └── audit/*.md # 独立巡检报告
├── .claude-plugin/ # Claude Code marketplace 注册
├── .codex-plugin/, .agents/ # Codex 注册
└── examples/, install.sh, install-codex.ps1
核心管线
x-spec ─→ x-req ─→ x-dev ─→ x-verify ─→ x-qa-gate ─→ x-fix
(docs/spec/) (task/) Gate ① Gate ② 批量修+增量复审
命令+smoke/e2e 风险路由：默认 RC 综合 / 高危 R1→R2→R3
x-qdev ─→ 定向验证 ─→ DoD 证据闭环 ─→ ✅
├─ Q2：一个综合 reviewer
└─ Q3：升级 x-req → x-dev 完整流程
- x-spec 产出
docs/spec/<spec-name>/独立需求包，docs/spec/README.md做索引，迭代原地更新 - x-req 产出
dev-pipeline/tasks/<task>/（README 含spec:字段指向归属 spec，一步到位） - x-plan 已废弃，功能合并到 x-req
独立巡检（不在主流程）：x-audit-perf / x-audit-style / x-audit-arch，由用户手动触发或里程碑后跑。x-audit-arch 聚焦架构一致性 + 单一事实源（结构性视角），与 x-audit-style（表层规范）、x-qa-gate R1（spec 正确性）不重叠，边界见 skills/x-audit-arch/SKILL.md。
skill 间契约（改 skill 前必读）
| 契约 | 内容 |
|---|---|
| spec 需求包目录 | x-spec 产出 docs/spec/<spec-name>/，docs/spec/README.md 汇总 spec 导航，spec 目录含 7 文件。x-req 的 README spec: 字段指向 docs/spec/<spec-name>，一个 task 只归属一个 spec |
dev-report.md schema |
x-dev 使用 skills/x-dev/templates/dev-report-template.md（含 risk: default/high 字段，Gate ② 路由依据）；x-verify 的必跑清单 = dev-report 命令表 + task README Smoke/E2E 用例（manual 用例列入待人工验收）；x-qdev 默认使用 skills/x-qdev/templates/dev-report.md，用户指定完整门禁时改用 x-dev schema |
| 测试分层契约 | x-spec 写验证策略；x-req README 显式列 smoke/e2e 验收用例；单元/契约/边界测试由 x-dev 按实际改动补齐，并写入 dev-report.md 验证命令清单 |
.fix-counter 共享 |
路径 dev-pipeline/tasks/<task>/reports/.fix-counter。语义 = 批量修轮数（一轮 = 一份 issue 清单的整体修复）。x-verify 首次创建，x-fix 按轮递增，x-qa-gate 在 Gate ② 最终 pass 后重置。3 轮上限，三方共享 |
| reviewer 子 agent | x-qa-gate 按 risk: 风险路由：默认线 dispatch 一个综合 reviewer RC（references/rc-unified.md），高危线串行 dispatch R1/R2/R3；初始 prompt 预算为 10,000 estimated tokens，使用 manifest + 路径 + diff 命令 + completeness gate |
| 一轮列全 | 所有 reviewer 穷尽列出全部问题候选后判定；每条返回 task、severity、loc、msg 和证据，并省略编号；回执包含覆盖声明、完整问题候选和穷尽声明；严重度 P0/P1/P2 唯一定义在 skills/x-qa-gate/SKILL.md |
| reviewer 只读 | RC/R1/R2/R3 只输出 review 回执；修改统一走 x-fix |
| issue 登记与状态写权 | 主 agent 逐条调用 python3 tools/xdev.py flag ... --json；本轮首条带 --new-round。代码分配 issue-<n>、写 ledger、把 P0/P1 task 降为 [!] 🔴；recovered:true 时主 agent 用原参数再次调用。reviewer、子 agent、x-fix 保持 ledger 与状态列原样 |
| x-fix 批量修 + 增量复审 | x-fix 一次修完一轮 issue 清单（P0 全修且各固化一条可复跑反例、P1 修或豁免、P2 登记），产出带 issue ID 的逐条处置表；复审尽量由同一个 reviewer 承接，只看 issue 处置 + fix 增量 diff，熔断条件见 skills/x-qa-gate/SKILL.md |
| 门禁回执 | x-verify / x-qa-gate / x-fix 每个节点结束在对话中输出统一回执（P0/P1/P2 计数 + 拦截来源维度 + 处置），零问题也输出；QA Gate issue ledger 由 flag 生成。格式见 skills/x-qa-gate/SKILL.md「回执与状态」 |
| x-cr 手动调查 | skills/x-cr/SKILL.md 是手动软件正确性调查入口，独立于自动门禁；流水线 gate 逻辑写到 x-qa-gate |
| 状态码 | ⏳ 未开始 / ▶️ 进行中 / 🟡 待验证 / 🔴 验证失败 / 🟢 证据通过 / ✅ 已完成 / ↗️ 已升级。x-dev 最多到 🟢，✅ 由完整 review 升级；x-qdev 按 Q0/Q1 主 agent 或 Q2 综合 reviewer 路线关闭，Q3 使用 ↗️ 并由 full task 负责最终完成 |
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.
- 5d ago First seen · 107 lines · 2,657 tokens per session scan A 70c146f24667
x-dev-pipeline CLAUDE.md is an instructions file published in the GitHub repository KtKID/x-dev-pipeline (12 stars, last pushed 1mo ago), licensed MIT. It adds 2,657 tokens to every session, about $0.0133 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
procoder tests.instructions.md
Instructions for azrtydxb/procoder, covering tests in procoder, every test carries its mutation, assertions, portability and fixtures.
lead-dev-os CLAUDE.md
Instructions for CaptainMe-AI/lead-dev-os, covering claude.md, project, terminology, plugin architecture and repository structure.
claude-code-ui-ux-skill CLAUDE.md
Claude Code instructions for nicohodt/claude-code-ui-ux-skill, covering claude.md, project overview, search command, architecture and sync rules.
team-of-agents CLAUDE.md
Instructions for pranav8494/team-of-agents, covering claude.md, project rules for ai assistants and git commits.
Myco AGENTS.md
Instructions for BreetyGreen/Myco, covering agents.md — orientation for ai agents & contributors, what this project is, architecture (two layers), single source of truth: engine/agents.json and key files.
claude-sdlc CLAUDE.md
Claude Code instructions for lantisprime/claude-sdlc, covering claude.md, what this repo is, design intent — read this before "improving" anything, eat your own dog food and hook strictness philosophy.