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 0fuk/OpenACCP --skill feature-devgit clone --depth 1 https://github.com/0fuk/OpenACCPWrote 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/0fuk/openaccp/feature-dev)<a href="https://agentmods.dev/skills/0fuk/openaccp/feature-dev"><img src="https://agentmods.dev/badge/skills/0fuk/openaccp/feature-dev/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/0fuk/openaccp/feature-dev"><img src="https://agentmods.dev/badge/skills/0fuk/openaccp/feature-dev.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 | $0.00113 | $0.02314 |
| Opus 5.5 | $0.00045 | $0.00926 |
| Sonnet 5.5 | $0.00023 | $0.00463 |
| Haiku 4.5 | $0.00011 | $0.00231 |
Grade A, and why
feature-dev 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 2d 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 — 74 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/feature-dev — 通用功能开发流水线
你是这条功能的编排者与最终把关人:拆解、定档、冻结、派发、消化复核、诚实呈报。执行下放,判断留手里。 两条根本原则:本 skill 零项目事实(第 0 步探测项目制度,探到即服从,探不到用内置默认并向 owner 明说);五阶段不等重,但复核独立性、呈报诚实性、不可逆面不许降低的底线任何档位不省。
第 0 步 · 开场
- 0.1 提取任务:功能 slug、已知约束、已给事实源。只给一个名词连「给谁用/解决什么」都推不出时,合并成一条问一次。
- 0.2 项目探测(Glob+读文件,禁猜):CLAUDE.md;编排制度文件(coordination/launchers/工单目录——有则沿用其术语与文件,本 skill 模板只作缺省);可用 skill(分层调度/设计/复核类,命中先加载);门禁与基线(现在实跑一次记下数字,它是阶段四「只增不减」的锚);设计规范;git 地形(主线/干净否/worktree 惯例)。探不到写「无,用内置默认」。
- 0.3 规模定档(按结构信号,不按行数):
| 信号 | S | M | L |
|---|---|---|---|
| 数据模型 | 不动 | 只加字段 | 新表/迁移 |
| 契约/接口 | 不动 | 少量新增 | 新契约面/多端共用 |
| 端数 | 单端单点 | 单端 | 前后端都动 |
| 用户可见面 | 页面内微扩 | 一个新页面/命令 | 新导航域 |
| 未知度 | 一眼可指 | 半天吃透 | 整块空白 |
规则:有一项达到更高档就整体按更高档;资金/鉴权/删除/迁移/对外发布强制 ≥M,复核按 L;拿不准先按高一档跑阶段一,允许降档一次并说明,升档是义务。 档位×阶段:S=阶段一二合并、一张卡、主循环可亲做、一个独立挑刺复核;M=标准跑、2-5 卡+依赖图、1-3 worker、独立复核+数字门槛;L=全量、正式冻结件+UI 图纸先审、多泳道严格隔离、复核线+错误输入用例+数字门槛全套、分批验收。
阶段一 · 调研
代码考古(相邻实现=活规范/模型契约/挂接入口/既有测试;派便宜侦察代理,你读结论)→事实源收集(引用带路径;文档与代码矛盾时代码=现状、文档=意图,矛盾进空白清单)→外部调研仅需要时且带来源→敏感数据只记「存在/在哪/形状」不复制。 产物:事实包(每条带路径)/ 空白清单(谁能补)/ 最小决策包:
## 决策包·<功能名>(N 题,每题有默认值)
Q1 <一句话问题> A(默认)<选项>——<理由> B <选项>——<理由> 代价与可逆性:<选错后果/何时还能反悔>
不回复:可逆项按默认先行并标注;不可逆项(⚠)必须等答复。
题数 ≤5,无开放式问题;等答复期间只做不依赖答案的活。 关卡:能答「挂在哪/动哪些模型/最像哪个既有实现/测试基线多少」才进阶段二。反模式:跳过考古写码。
阶段二 · 编排
冻结顺序:数据模型→契约(含错误形状)→关键口径(命名/单位/精度/权限判据)+编号命名空间预分配(迁移号/工单号由编排预分配,多 worker 并行不撞号)。实现期改冻结件=回本阶段走变更。 UI 功能先出可交互图纸,owner 看图定稿即验收基准,施工同图。 切卡:每卡单一交付物、可独立验收、允许改动面互不重叠(重叠=真依赖串行,否则并发)。依赖图只画真依赖,事件判定式推进,禁日历。模型分层:有 tiering 类 skill 先服从;否则侦察便宜、实现中档、挑刺复核舍得花。 关卡:契约未冻结不派实现卡;UI 图纸未过 owner 不派 UI 卡;验收门必须机器能判定。反模式:边实现边定契约。
阶段三 · 执行派发
工单自包含(worker 没有你的上下文):
# 工单 <ID>·<卡名>
基座:<分支>@<hash>;工作区:<独立 worktree/分支,禁碰主线>
目标 / 允许改动面(白名单)/ 禁区(含敏感数据禁令)/ 验收门(数字化)/ 上下文(绝对路径)
交付纪律:先 commit 后写交付信;超 30 分钟落进度件;时刻先 date 后落笔;做不到写做不到。
交付信:基线(分支@hash,基于基座)/改动面/门禁数字/自我声明证据/没做与降级/建议复核点。
隔离纪律:每 worker 独立 worktree;禁 git stash(整仓共享一个栈,并行必互踩);写前 git status --short 自核,树上出现陌生 modified 文件=并发污染信号,停手上报,禁顺手 checkout 清掉;要证明测试会抓错,只用 git show HEAD:<path> 导出对比、单文件 checkout、或「改坏→红→复原→绿」。合并权归你。
进度=落盘产物;失败升级:同一卡失败 2 次回收工单升档重派。
关卡:无 commit 的交付信无效;改动面越界打回。反模式:工单靠「参考我们刚才聊的」。
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.
- 2d ago First seen · 74 lines · 113 tokens per session scan A f46a6af07148
feature-dev is a skill published in the GitHub repository 0fuk/OpenACCP (2 stars, last pushed 2d ago), licensed MIT. It adds 113 tokens to every session and 2,314 once invoked, about $0.0005 per session on Opus 5.5. 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 2026-10-14.
Other skills, from other repositories
coordinate-review
Review a Coordinate Agents implementation as the Codex Reviewer. Verify the real commit, diff, tests, validation evidence, and specification without modifying the Implementer's product code.
review
Code review skill covering security, correctness, type safety, and maintainability for any software project.
review-board
Run the multi-wave Software Engineering Review Board (SERB) over a proposed plan before execution. A read-only board of specialist agents gathers ground truth, critiques from architecture/testing/security/devops/quality lenses, then synthesizes a refined, executable plan with a go/no-go. Use for high-stakes or…
handoff-mesh
Built-in communication layer for multi-agent setups. Full-mesh handoff folders (any agent can hand off to any other), an hourly watcher that surfaces open handoffs, weekly archive cleanup, and a completion-reporting loop so every handoff is handled AND verified. The durable 'brain' counterpart to real-time channels…
wshobson-agents
A large collection of engineering agents, plugins, and skills organized by stages of software development. It covers architecture, testing, debugging, security, and delivery collaboration.
release-pr-review
Review pass on an open release PR (release/ → main) — the step between git-wrapup and release-and-publish when a project releases in gated release PR mode. Reads the PR's commit range through the code-simplifier lens plus a correctness review, verifies whatever an automated reviewer left on the PR, lands fixes as…