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 skills/lzzlhy/archview/skillnpx skills add LZZLHY/archview --skill skillgit clone --depth 1 https://github.com/LZZLHY/archviewWrote 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/lzzlhy/archview/skill)<a href="https://agentmods.dev/skills/lzzlhy/archview/skill"><img src="https://agentmods.dev/badge/skills/lzzlhy/archview/skill.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 | $0.00055 | $0.07194 |
| Opus 5 | $0.00028 | $0.03597 |
| Sonnet 5 | $0.00011 | $0.01439 |
| Haiku 4.5 | $0.00006 | $0.00719 |
Grade A, and why
archview 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 — 436 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/archview
给指定工作区的代码节点写语义摘要,让 ArchView 面板从「一堆文件名」变成「能看懂的架构」。
这份提示词的结构参照 Understand-Anything 的
skills/understand/SKILL.md(MIT,见仓库 NOTICE),流程换成 ArchView 的。和 UA 最根本的一处不同:UA 让 LLM 同时挖结构又写语义,所以它需要一堆修补脚本 擦屁股(规范化 ID、丢悬空边、翻转方向搞反的边)。ArchView 里结构是 CodeGraph 给的 确定性事实,你只写语义。因此本流程里没有任何一步让你产出节点或边 —— 如果你发现自己在写 JSON 里的
source/target/type,你已经跑偏了。
选项
$ARGUMENTS 可以包含:
- 一个目录路径 —— 目标工作区根(默认当前工作目录)
--module <key>—— 只处理这一个模块(模块 key 见 AGENT-GUIDE.md 的模块表)--stale-only—— 只重写 hash 失配的过期摘要,不碰缺失项--limit <n>—— 本轮最多写 n 条(大工作区分批推进用)--dry-run—— 只报告要写哪些节点,不落盘
铁律(读完再动手)
铁律 1 —— 你永远不写拓扑
节点和边只能由 CodeGraph 的 tree-sitter 输出派生。你能提供的字段只有两个:
summary 与 tags。
具体禁止:
- 绝不允许手写、修改、"修正"
<workspace>/.archview/graph.json。它是派生物, 重建会整份重写。你的编辑不但会消失,在消失之前还会让面板显示假架构。 - 不要动
.archview/meta.json、.archview/briefs/—— 同样是派生物。 - 不要在摘要里把推断出来的 import / 调用 / 继承当事实写。这些关系简报里已经有了
(
importsFrom/importedBy/renders/renderedBy),直接引用。 - 发现拓扑不对(缺边、方向反、模块归属错),报告它,不要动手改图。 那是 builder 或 deriver 的 bug,改图只会掩盖它。
铁律 2 —— 每个节点的 summary 必须非空
没有 LLM 摘要时系统会填确定性兜底句(<名字> —— <路径> 中的 <kind>)。
那不是摘要,它只是为了不让校验器报警。简报里 summarySource 不是 "llm"
的每一条都是你的活(真实取值见阶段 2 那张表)。
铁律 3 —— 摘要的 key 必须是简报里的 nodeId 原文
节点 ID 方案是冻结的,摘要文件用它做 key。拼错的 key 不会报错,只会静默变成 孤儿摘要,在重建时被丢掉。复制,不要重新拼装。
进度汇报
每个阶段开始时打一行状态,写完一批报一次。工作区大的时候这一步不是礼貌,是让人知道 你还在动。
[阶段 N/6] <阶段名>…例:
[阶段 3/6] 写 entry 模块的摘要(83 个文件节点,缺 12 条)…
阶段 0 —— 摸清现场
-
确定
WORKSPACE:$ARGUMENTS里第一个不以--开头的参数就是它;没有就用当前 工作目录。确认<WORKSPACE>/.archview/存在。 -
读
<WORKSPACE>/.archview/AGENT-GUIDE.md。这是唯一的必读文件。它是生成物,装着这个工作区此刻的实况:项目名与描述、模块清单(各自文件数)、 检测到的语言、摘要覆盖率、具体缺哪些节点、简报文件的确切路径、该读哪几份语言/ 框架指导、输出格式、MCP 配置片段、自检清单。
- 文件不存在时,在 ArchView 仓库目录里跑
pnpm exec archview-skill guide --workspace <WORKSPACE> --write(或node <仓库>/archview/packages/skill/dist/bin/skill.js guide --workspace <WORKSPACE> --write)生成。 不要用npx archview-skill—— 这个包还没发布到 npm,npx会去 registry 找它然后失败; 裸命令archview-skill也不在 PATH 里。跑不了就退回读<WORKSPACE>/.archview/briefs/自己判断(信息少一些,但够干活)。 - 文件在,但里面的数字看着与
.archview/的现状矛盾(比如它说缺 40 条、而 summaries 里已经有了),说明它是旧的:重新生成一次,别拿旧数字干活。
- 文件不存在时,在 ArchView 仓库目录里跑
What ships with it
60 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- frameworks/django.md 7.4 KB
- frameworks/express.md 6.8 KB
- frameworks/fastapi.md 7.2 KB
- frameworks/flask.md 6.9 KB
- frameworks/gin.md 7.4 KB
- frameworks/nextjs.md 7.6 KB
- frameworks/rails.md 7.9 KB
- frameworks/react.md 7.4 KB
- frameworks/spring.md 8.0 KB
- frameworks/vue.md 8.1 KB
- languages/arkts.md 9.1 KB
- languages/cfquery.md 7.0 KB
- languages/cfscript.md 7.1 KB
- languages/cobol.md 6.8 KB
- languages/cpp.md 4.2 KB
- languages/csharp.md 3.9 KB
- languages/css.md 3.3 KB
- languages/dart.md 5.5 KB
- languages/dockerfile.md 4.3 KB
- languages/erlang.md 6.3 KB
- languages/go.md 4.3 KB
- languages/graphql.md 4.3 KB
- languages/html.md 4.3 KB
- languages/java.md 4.3 KB
- languages/javascript.md 5.1 KB
- languages/json.md 5.2 KB
- languages/kotlin.md 5.4 KB
- languages/lua.md 6.5 KB
- languages/luau.md 6.7 KB
- languages/markdown.md 5.4 KB
- languages/nix.md 6.9 KB
- languages/objc.md 6.3 KB
- languages/pascal.md 6.7 KB
- languages/php.md 5.8 KB
- languages/protobuf.md 4.4 KB
- languages/python.md 5.0 KB
- languages/r.md 6.6 KB
- languages/ruby.md 5.2 KB
- languages/rust.md 5.3 KB
- languages/scala.md 5.4 KB
- languages/shell.md 5.2 KB
- languages/solidity.md 7.5 KB
- languages/sql.md 5.3 KB
- languages/swift.md 5.7 KB
- languages/terraform.md 5.8 KB
- languages/typescript.md 5.9 KB
- languages/vbnet.md 6.8 KB
- languages/yaml.md 5.8 KB
- package.json 2.1 KB
- README.md 4.9 KB
- scripts/pack.mjs 2.7 KB runs code
- src/assets.ts 6.7 KB runs code
- src/bin/skill.ts 12 KB runs code
- src/check.ts 17 KB runs code
- src/facts.ts 11 KB runs code
- src/guide.ts 72 KB runs code
- src/hosts.ts 6.8 KB runs code
- src/index.ts 1.8 KB runs code
- src/install.ts 25 KB runs code
- src/seam.ts 1.8 KB runs code
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 · 436 lines · 55 tokens per session scan A 00408a7967c2
archview is a skill published in the GitHub repository LZZLHY/archview (0 stars, last pushed 10d ago), licensed MIT. It adds 55 tokens to every session and 7,194 once invoked, about $0.0003 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 2026-08-31.
Other skills, from other repositories
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.
brainstorming
You MUST use this before any creative work - creating features, building components, adding functionality, or modifying behavior. Explores user intent, requirements and design before implementation.
auto-perf-optimize
Run agent-driven VS Code performance or memory investigations. Use when asked to launch Code OSS, automate a VS Code scenario, run the Chat memory smoke runner, capture renderer heap snapshots, take workflow screenshots, compare run summaries, or drive a repeatable scenario before heap-snapshot analysis.
chat-perf
Run chat perf benchmarks and memory leak checks against the local dev build or any published VS Code version. Use when investigating chat rendering regressions, validating perf-sensitive changes to chat UI, or checking for memory leaks in the chat response pipeline.
chat-pet-sprite-creation
Use when creating or changing VS Code chat pet sprite art, sprite sheets, state animations, eye treatments, Stable/Insiders variants, or pet transitions under src/vs/workbench/contrib/chat/browser/widget/media/chatPet.
cpu-profile-analysis
Analyze V8/Chrome CPU profiles (.cpuprofile) and DevTools trace files (Trace-.json). Use when: profiling performance, investigating slow functions, comparing code paths, finding bottlenecks, analyzing timeToRequest, understanding call trees from sampling profiler data, analyzing layout/paint/rendering, investigating…