A Claude Code plugin containing a full software-development workflow from product definition through implementation, testing, review, release, and learning. Its steps produce files that scripts can check or require a human approval.

Use it to turn requests into specifications, contracts, plans, implementations, acceptance checks, pull requests, releases, and retrospective records.

It makes requirements and completion rules explicit before code is changed. This reduces untracked decisions, unchecked failures, and confusion about whether work is actually done.

ai-dlc — Plugin for Claude Code by XRenSiu

Install

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.

Clone the repo

git clone --depth 1 https://github.com/XRenSiu/claude-code-forge

Made for: Claude Code.

Wrote 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.

README.md

[![agentmods](https://agentmods.dev/badge/plugins/xrensiu/claude-code-forge/ai-dlc/github.svg)](https://agentmods.dev/plugins/xrensiu/claude-code-forge/ai-dlc)

Your own site

<a href="https://agentmods.dev/plugins/xrensiu/claude-code-forge/ai-dlc"><img src="https://agentmods.dev/badge/plugins/xrensiu/claude-code-forge/ai-dlc/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.

Your own site · 80×15

<a href="https://agentmods.dev/plugins/xrensiu/claude-code-forge/ai-dlc"><img src="https://agentmods.dev/badge/plugins/xrensiu/claude-code-forge/ai-dlc.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session not measured What this adds to a session before it is invoked.

When invoked not measured The manifest loads nothing itself; its 28 skills, 5 agents cost 8,523 tokens a session between them.

Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.

Origin original No closer match found in the catalogue.

Security

Grade A, and why

ai-dlc 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 today.

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.

plugins/ai-dlc/.claude-plugin/plugin.json · 92 lines

How it starts

The opening of the file, as written. The whole thing — 92 lines — stays where its author put it; the contents beside it link to each section on GitHub.

{
  "name": "ai-dlc",
  "description": "AI-DLC — 一条产研自闭环的软件开发流水线，做成 Claude Code 插件：世界 → 本体 → 契约 → 标准 → 计划 → 实现 → 验收 → 交付 → 学习，每一步的产物要么能被脚本检，要么被一道只能人签的门挡住，没有第三种。28 个 skill 组织成九环 + 一根脊柱（docs/ARCHITECTURE.md）。\n\n三条主线。**上半段建世界**：/psl 把体验性需求写成六层 Product Specification Language，/psl-derive 推出 DOS 提案 / workflow / 形态草案（每条决策引用一个 PSL-ID）并用 N 次隔离推导的分歧集当 G1 议程。**下半段收敛交付**：/issue → /donewhen-extract（或 /acceptance-spec）→ G2 哈希冻结判据 → /plan-cards 切成自包含任务卡 → /implement 隔离实现 + /commit 三道闸 → /acceptance-fleet 并行派发六个审查 skill → /meta-judge 合成四态棘轮 → /pr → /review-loop（评论当待验证主张）→ /release → 归档。**横切**：/dos-extract 与 /invariant-extract 管本体与不变量，routing.yaml 管失败按层回流与分层预算，/retro 与 /tune 管度量与环参数。\n\n设计上不肯让步的几条。**路由确定化，质量交给模型**：状态机、路由表、执行图、六个环契约全是数据，由脚本强制，引擎只提议——`aidlc_state.py advance` 对着状态检前置条件，不对着模型的说法检。**判据先于代码且冻结**：done_when.yaml v2 是唯一契约 schema，G2 签它、L5 再签一次测试，之后改锁内文件必须附变更提案。**评估者与被评估者分离**：实现者只见卡 + AC 子集 + 红基线，看不到评审判据与隐藏集。**声明了但没求值 ≠ 通过**：分析器跑不了退出 3，审查没跑完记为 unevaluated，两者都不许报绿。**失败归层不原地重试**：同指纹两次、震荡、平台期分别路由到方案层或任务层，预算耗尽写失败报告交人。**账本只增不删**：产物可回滚，判据与失败记录不回滚。**三道门只能人签**：G1 世界裁决 / G2 判据冻结 / G3 例外复核，`--autopilot` 与自治阶梯都不代签。\n\nv1.0.0 起流程按三个正交旋钮伸缩（借鉴 AWS AI-DLC 2.0 的 scope grid，见 docs/reports/aidlc-gap-2026-09-07.md）：**广度**是 sizing.yaml 的阶段网格，由 next_allowed / prereqs 强制、verify_sizing.py 九条 lint 核验，never_skippable 把三道门挡在旋钮之外，跳过要用可数的证据换、每次写一条带理由的账本行；**测试量**是下界，derive_counts.py --strategy 检，低于地板退出 4；**深度**只是声明，因为没有脚本能判「这份文档够不够细」——文档如实标注，不假装它是闸。同版新增解释日记（notes.md 四格：Interpretations / Deviations / Tradeoffs / Open questions，门禁前逐字呈现不筛选，Open questions 不晋升，晋升的规则下一次 init 才编译进去）、审查完成度闸 verify_review_complete.py、跨制品术语传感器 verify_vocabulary.py（B 档）、自治阶梯 autonomy、启动前算有效规模的 plan、装置健康度 doctor。\n\n证据等级诚实登记：全部 skill 为 static_only —— 结构过审 + 脚本在 fixture 上冒烟（条数以 `eval/smoke.sh` 结尾行为准，不手抄；关键行为配变异证明），**带 / 不带这套纪律的行为层对照只跑过一轮十题**，且出题人与被测者同源。能说什么、不能说什么写在 docs/evaluation.md。\n\nv1.6.0（2026-09-12）按「它是不是一条闭环」重读一遍，修了七处声明与代码不一致：逃逸缺陷有了登记入口（`aidlc_state.py escape`：计到归因层、追加 escape-defects.md、镜像进归档——此前没有任何命令写 escape 事件，复盘的逃逸率结构上永远是

Read the full file on GitHub · 92 lines

Contents

What it installs

The manifest is a name and a version. 28 skills, 5 agents travel with it, and installing the plugin installs all of them — 8,523 tokens a session between them. Each is measured on its own page, and each can be installed alone.

Changes

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.

today Changed c97e95b70d4e
yesterday Changed d999b39031d6
2d ago Changed 0fd346700e81
3d ago First seen · 92 lines scan A a831b2ea0277

Subscribe to this mod's changes

ai-dlc is a plugin published in the GitHub repository XRenSiu/claude-code-forge (2 stars, last pushed today), licensed MIT. Its token cost is not measured: this kind of file is read by the harness, not the model. 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 2026-09-09.

Planning Testing

load-tester

Load and stress testing for APIs and web services.

rohitg00/awesome-claude-code-toolkit · not measured

kit

Spec-driven Claude Code workflow with verification pipeline. Hooks, slash commands, and verification subagents.

dwarvesf/dwarves-kit · not measured

looptimal

Turns an objective into a delivered, VERIFIED OUTCOME: frames a hash-pinned sealed acceptance suite, designs the right loop (the loop-design wizard, formerly LoopPrint), war-games it forward, executes with dynamic domain-expert sub-agents (maker != checker), and gates completion on a separate verifier re-running the…

Renn-Labs/Looptimal · not measured

pict-test-designer

Design comprehensive test cases using PICT (Pairwise Independent Combinatorial Testing) for any piece of requirements or code. Analyzes inputs, generates PICT models with parameters, values, and constraints for valid scenarios using pairwise testing. Outputs the PICT model, markdown table of test cases, and expected…

eric861129/SKILLS_All-in-one · not measured

local-test-plugin

Symlink-based local plugin testing for development workflow.

devstefancho/skills · not measured

headless

Headless browser automation for site comparison, E2E testing, and anti-bot-aware web research.

paddo/claude-tools · not measured