DESIGN.md@dsh-plugins/dsh-sleuth-copilotA

DESIGN.md@dsh-plugins/dsh-sleuth-copilot from pillumina/ascend-sleuth is a DESIGN.md file: the design rules coding agents follow in that repo.

Install it, pinned
mdr add pillumina/ascend-sleuth/dsh-plugins/dsh-sleuth-copilot/DESIGN.md@git:20261105.43c4f37

Pins git:20261105.43c4f37 in mdr.lock. Pin the fingerprint instead.

Audit grade
A, 14 of 14 checks passed
Kind
DESIGN.md file
Repository
pillumina/ascend-sleuth
Path
dsh-plugins/dsh-sleuth-copilot/DESIGN.md
Versions recorded
5
Latest version
git:20261105.43c4f37, committed Nov 5, 2026
First version recorded
Nov 4, 2026
Size
17,212 bytes
Repository stars
5
License
MIT

Collected from public GitHub. Every version is saved exactly as it was and never changes. Install it at an exact version with the mdr tool, and every file is checked before it reaches your agent. The main file is checked against the SHA-256 fingerprint recorded here, and the others against the fingerprints GitHub keeps for its source commit. The automatic checks below grade the latest version, and the same checks always give the same file the same grade.

What the file says

# 诊断交互层:插件点地图与取舍

这份文件记录「为什么是这几个面」与「还有哪些点可以考虑」。实现清单只看
[README.md](README.md);这里放判断依据。

## 读者与场景

- 读者:现场工程师与技术支持,白天连续处理多单;一只眼看日志,一只眼看着对话。
- 任务循环:收集材料 → 判断 → 给结论 → 回报 / 沉淀。
- 一次中断的代价:切到别的界面、找到那一单、再切回来。三分钟以上的诊断会被切走十几次。

## 六条交互原则

1. **不打断**:读状态与常见写入都在对话内完成。
2. **渐进披露**:默认一行;展开才给细节;偏好在下次打开时保留。
3. **下一步只有一个**:主按钮永远指向当前最该做的事,其余进「更多操作」。
4. **写入可核对**:写前给出对象与字段,写后给结果;并发冲突给原因与刷新入口。
5. **一处定义、多处一致**:三个面共用同一份数据、同一套动作措辞与同一份偏好。
6. **降级可读**:没数据、没权限、写不了时给出原因与下一步,不拿别处的数据顶上。

## 旅程 × 插件点

「已做」是本包已实现的面;「评估过」是看过扩展点后没做的,附理由。

| 时刻 | 摩擦 | 插件点 | 状态 |
|---|---|---|---|
| 开单 | 不记得从哪开始 | `conversation.hero.options`(新会话行加「诊断」入口) | 评估过:`/diagnose` 已在空会话首屏够用,加一个按钮收益低 |
| 诊断中 | 不知道到哪步、在等什么 | `conversation.input.dock` 状态条 | 已做 |
| 诊断中 | 信息块挡消息 | 状态条收起 + 偏好持久化 | 已做 |
| 诊断中 | 要回报 / 收尾却要开面板 | 状态条「更多操作」+ 总览写入 | 已做(回报轴仅生成指令,见 README) |
| 诊断中 | 回报结果该写什么 | 回报表单(结果 / 来源 / 备注 + 指令预览) | 已做 |
| 诊断中 | 想查 case 上游有没有新结论 | 「查 case 上游状态」生成指令 | 已做:菜单先显示 gh 就绪状态(与 doctor 同一件事),prompt 让 agent 走 `verify_upstream_outcomes.py` 幂等落账(同一结局不重复累加) |
| 诊断中 | 想把结论发给同事 | 一键复制摘要(markdown 纯文本) | 已做 |
| 诊断中 | 想跳到该诊断的会话 | `sessions` / `uiWorkspace` 的打开会话方法 | 已做,仅轨迹写了 `harness_session` 时出现 |
…

Read the whole file at its exact version.

How to install

The exact contents, by fingerprint
mdr add pillumina/ascend-sleuth/dsh-plugins/dsh-sleuth-copilot/DESIGN.md@sha256:ff1255f383d4753a

The line at the top of this page pins the label git:20261105.43c4f37, which is what most people want. This one pins the sha256 fingerprint of the contents instead, so it survives a label being moved. Either way the exact version is written to mdr.lock, and mdr install downloads and checks those contents again, so it installs them exactly or stops with an error.

Badge

mdr badge

[![mdr](https://markdownregistry.com/badge/art_kral2evhbijybmae.svg)](https://markdownregistry.com/a/art_kral2evhbijybmae)

2 badge image fetches in 30 days. What the badge shows.

Versions

The registry has recorded 5 versions of this file, between Nov 4, 2026 and Nov 5, 2026, more than one a day.

versioncommittedcommitsizeaudit
git:20261105.43c4f37 latest2026-11-05 43c4f37 17,212 BA view · diff
git:20261104.198f32a2026-11-04 198f32a 16,470 BA view · diff
git:20261104.16db2292026-11-04 16db229 15,152 BA view · diff
git:20261104.a5dfeab2026-11-04 a5dfeab 14,948 BA view · diff
git:20261104.cbb76532026-11-04 cbb7653 15,051 BA view

Audit of the latest version

A  14 of 14 checks passed. Fixed automatic checks with no AI involved, so the same checks give the same answer every run.
  • pass: Size between 200 bytes and 200 KB (17212 bytes)
  • pass: No zero-width or bidi control characters
  • pass: No instruction hidden inside an HTML comment
  • pass: No link to an exfiltration or paste host
  • pass: No credential-shaped string
  • pass: No instruction to send local credentials anywhere
  • pass: No text hidden with inline styles
  • pass: No prompt-injection phrasing
  • pass: No curl or wget piped into a shell
  • pass: No recursive delete of root, home or parent
  • pass: No instruction to read or print local credentials
  • pass: No base64 blob over 200 characters
  • pass: No link to a raw IP address
  • pass: No script tag

Source

GitHub

pillumina/ascend-sleuth · 5 stars · license MIT · pushed 2026-11-06 · branch main

API

GET https://markdownregistry.com/api/v1/artifacts/art_kral2evhbijybmae
GET https://markdownregistry.com/api/v1/resolve?ref=pillumina/ascend-sleuth/dsh-plugins/dsh-sleuth-copilot/DESIGN.md
GET https://markdownregistry.com/api/v1/blob/ff1255f383d4753ad6677b67c8bb6216f3cca3d998c991ed415abe4af9d5e3b2

Your AI assistant does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.

Guides: what DESIGN.md is and how agents use it · which agent file does what.

More from pillumina/ascend-sleuth

AGENTS.md AGENTS.md file
pillumina/ascend-sleuth · AGENTS.md
git:20261102.0c3a557 · audit A · 5 stars
CLAUDE.md CLAUDE.md file
pillumina/ascend-sleuth · CLAUDE.md
git:20261102.0c3a557 · audit A · 5 stars
ask SKILL.md skill
pillumina/ascend-sleuth · skills/ask/SKILL.md · 昇腾知识库快问快答:用户问一个能一句话问完的问题(错误码或环境变量取值、某类问题怎么定位、 本次现象先查什么、某个术语是什么、库里有没有收录)时,用单个索引文件定位,读不超过 2 个词条, 按固定 5 行格式答「一句话 + 表或候选 +…
git:20261030.a103b40 · audit A · 5 stars
diagnose SKILL.md skill
pillumina/ascend-sleuth · skills/diagnose/SKILL.md · 昇腾训练/推理问题的核心诊断循环:收集症状、按 triage-tree 路由、两阶段加载 并验证 Tier 2 case、命中给 fix(高危 root cause 改提示 halt)或转深度排查。 Tier-2…
git:20261104.0a4ba96 · audit A · 5 stars
doc-standards SKILL.md skill
pillumina/ascend-sleuth · skills/doc-standards/SKILL.md · 写或改任何给人读的文档之前,先定类型、再套骨架、最后过反例清单。覆盖 docs/ 六层、 README、定位报告、case/reference 词条、EV 卡、postmortem、PR body、skill 正文。给出九类…
git:20261009.3640a77 · audit A · 5 stars
doctor SKILL.md skill
pillumina/ascend-sleuth · skills/doctor/SKILL.md · 本机环境体检:一条命令问清这台机器上哪些流程能跑、哪些会降级(缺 Python 包、git、node、gh, traces/ 能否解析,知识库索引在不在),并给出每项缺失的修法。任务开始前环境可疑、换机器、 内网新部署、或流程中途以一句…
git:20261027.748c4a1 · audit A · 5 stars
evolve-check SKILL.md skill
pillumina/ascend-sleuth · skills/evolve-check/SKILL.md · 任何内容流程收尾的伴随演进评估(数据驱动的 self-evolving,非独立目标轮)。 用户下内容目标(补 case / reference 沉淀 / 诊断 / 拉取 / replay)后,执行中产生…
git:20261030.875bdba · audit A · 5 stars
issue-ingest SKILL.md skill
pillumina/ascend-sleuth · skills/issue-ingest/SKILL.md · 从外部 issue 源(GitHub 等)批量导入案例知识。编排上游 issue 拉取、硬过滤、价值评估、沉淀与状态标记:拉取精简元数据(fetch_issues.py,不含…
git:20261030.0194eb5 · audit B · 5 stars
knowledge-groom SKILL.md skill
pillumina/ascend-sleuth · skills/knowledge-groom/SKILL.md · 昇腾知识库的周期性维护引擎。先同步远端知识到确定 commit,用有界子 agent 比对本地草稿与现有 case 的语义(root cause / 机制 / 修复 / 证据),由确定性脚本校验并保守聚合,更新本地草稿评估状态; 再批处理…
git:20261105.abfd57d · audit A · 5 stars
preload-panel SKILL.md skill
pillumina/ascend-sleuth · skills/preload-panel/SKILL.md · 在 DSH 会话中加载 ascend-sleuth 面板,并处理「装了但 tab 不出现」。先按本机给模型的工具与桌面版 自带的东西选装载路:有 cordis_define / cordis_run 在会话内热加载;有…
git:20261103.12b1f39 · audit A · 5 stars
reference-ingest SKILL.md skill
pillumina/ascend-sleuth · skills/reference-ingest/SKILL.md · 从 GitCode 等可按 API 枚举的**文档仓**批量摄取官方文档,沉淀为 reference 先验词条(面向训练/推理的中断、精度、性能定位)。编排「仓库文档面普查 → 增量扫描(path×blob sha…
git:20261009.3640a77 · audit A · 5 stars
resume-diagnosis SKILL.md skill
pillumina/ascend-sleuth · skills/resume-diagnosis/SKILL.md · 续接一个被打断的昇腾诊断 session;也用于接手另一台机器打包过来的诊断(交接包)。 读活跃的 traces/*.yaml(每个并发诊断一个文件,含 trace),复述上次停在哪一步、排除了哪些 case、当前 active…
git:20261105.2f1c783 · audit A · 5 stars

Every file in pillumina/ascend-sleuth

Browse by kind, by grade A, or by owner.