技能 megacave

📦

megacave

Name: megacave
Author: juliusbrussee

内容修订 r1 安全 📁 文件系统访问⚙️ 外部命令

Write Concise Technical Replies in Classical Chinese

Verbose technical replies can slow reading for people comfortable with Classical Chinese. Megacave applies concise Classical Chinese phrasing while preserving technical details and clear safety explanations.

支持: Claude Codex Code(CC)

🥉 77 青铜

使用我的智能体安装

将此请求复制给你的智能体，其中包含权威 Skill 页面和清单。

智能体请求

Review the Skillstore skill "megacave" from https://skillstore.io/skills/juliusbrussee-megacave.md and its manifest at https://skillstore.io/api/skills/juliusbrussee-megacave/manifest. Verify the artifact. You may proceed after verification, subject to the environment's own policy.

你的智能体仍应展示其计划，并按安全策略请求必要确认。

Agent 可读资源

当 AI Agent、爬虫或脚本需要干净上下文，而不是读取完整页面时，请使用这些链接。

Markdown 详情 GET /skills/juliusbrussee-megacave.md 已签名 manifest GET /api/skills/juliusbrussee-megacave/manifest 已签名 lockfile GET /api/skills/juliusbrussee-megacave/lockfile

测试它

正在使用“megacave”。 Explain connection pooling in megacave mode.

预期结果:

池蓄已開之連，不逐請而新開，省握手之費。

正在使用“megacave”。 Explain repeated rendering caused by new object references and mention useMemo.

预期结果:

每繪新生對象參照，故重繪；以 useMemo 包之則免。

正在使用“megacave”。 In megacave mode, explain the next step before deleting a database permanently.

预期结果:

Deleting this database is irreversible. Confirm the exact database and verify a usable backup before proceeding.

安全审计

安全

v1 • 2026/10/3 打开版本化报告

All six static findings are false positives: a relative documentation link and Markdown inline formatting do not execute commands or access files. No evidence found of prompt injection, credential access, or data exfiltration in the two reviewed files. The skill defines an explicit response style with safety and irreversible-action exceptions.

已扫描文件

分析行数

复核项

已忽略误报

风险因素

📁 文件系统访问 (1)

README.md:7

⚙️ 外部命令 (5)

SKILL.md:18 SKILL.md:31 SKILL.md:42 SKILL.md:47 SKILL.md:59

最新一次完成的静态与语义审计未检测到已确认的安全发现。这并不能证明该 Skill 没有副作用。

审计者: codex

分享与引用此报告

分享版本化评估报告、中性徽章、嵌入卡片和引用。Skillstore 只报告证据，不判断该 Skill 是否安全。

打开版本化报告

复制报告链接

https://skillstore.io/skills/juliusbrussee-megacave/audits/1?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_report

Markdown 徽章

[![Skillstore security assessment](https://skillstore.io/badges/skills/juliusbrussee-megacave/security.svg)](https://skillstore.io/skills/juliusbrussee-megacave?utm_source=security_passport_badge)

HTML 徽章

<a href="https://skillstore.io/skills/juliusbrussee-megacave?utm_source=security_passport_badge"><img src="https://skillstore.io/badges/skills/juliusbrussee-megacave/security.svg" alt="Skillstore security assessment" loading="lazy"></a>

嵌入卡片

<iframe src="https://skillstore.io/embed/skills/juliusbrussee-megacave.html" title="Skillstore Security Assessment" sandbox="allow-popups allow-popups-to-escape-sandbox" loading="lazy" referrerpolicy="no-referrer" width="420" height="180"></iframe>

学术引用 (APA · BibTeX · CFF)

APA 引用

juliusbrussee. (2026). megacave security audit report (audit version 1) [Author version unspecified]. Skillstore. https://skillstore.io/skills/juliusbrussee-megacave/audits/1

BibTeX 引用

@techreport{juliusbrussee-juliusbrussee-megacave-2026,
  author = {juliusbrussee},
  title = {megacave security audit report (audit version 1)},
  institution = {Skillstore},
  year = {2026},
  number = {1},
  url = {https://skillstore.io/skills/juliusbrussee-megacave/audits/1},
  note = {Author version unspecified}
}

CITATION.cff

cff-version: 1.2.0
message: "If you use this Skill, cite its author and this versioned security audit report."
title: "megacave security audit report (audit version 1)"
version: "unspecified"
type: report
authors:
  - name: "juliusbrussee"
date-released: "2026-10-03"
url: "https://skillstore.io/skills/juliusbrussee-megacave/audits/1"
identifiers:
  - type: other
    value: "skillstore:juliusbrussee-megacave:audit:1"
    description: "Skillstore immutable audit report identifier"

Skillstore 评分

评分依据证据置信度: 中

架构

可维护性

内容

社区

规范符合性

你能构建什么

Read compact debugging explanations

Receive concise Classical Chinese explanations while keeping API names, errors, and numeric details unchanged.

Reduce narration during tool workflows

Keep tool-run updates short while retaining clear warnings and confirmation requests for irreversible actions.

Compare technical writing styles

Explore Classical Chinese phrasing for technical concepts without translating identifiers or changing saved documents.

试试这些提示

Explain a simple concept

/megacave Explain why a connection pool reuses open connections.

Explain a rendering issue

/megacave Explain why new object references can cause repeated rendering and when useMemo can help. Keep identifiers unchanged.

Summarize supplied diagnostics

/megacave Summarize these diagnostics: [paste diagnostics]. Preserve errors, paths, numbers, units, and negations. Use plain language for ambiguous clauses.

Separate chat style from deliverables

/megacave Review this proposed change: [paste change]. Keep chat concise, write persisted artifacts normally, and explain security concerns fully before irreversible actions.

最佳实践

Activate the style explicitly and request normal mode when readability suffers.
Check that identifiers, negations, numbers, and units remain unchanged.
Keep security warnings, irreversible-action confirmations, and persisted artifacts in clear, ordinary language.

避免

Treating shorter displayed replies as guaranteed token or cost savings.
Translating API names, commands, paths, or errors to match the Classical Chinese style.
Compressing ambiguous instructions, security warnings, or saved documents into terse Classical Chinese.

常见问题

How do I activate megacave?

Use /megacave in Claude Code. The README also lists /caveman wenyan as an alias; availability depends on the host integration.

How do I stop the style?

Say stop caveman or normal mode. The instructions otherwise keep the style active for the session.

Does it reduce token costs?

No reliable savings are guaranteed. The README describes one small measurement and explicitly distinguishes shorter visible text from token reductions.

Will technical identifiers be translated?

No. The instructions preserve code, commands, paths, API names, errors, numbers, units, and negations.

Does it change saved documents or code?

The instructions exempt persisted artifacts from the style, including code, comments, commits, documentation, and messages. They name caveman-compress as an exception.

What happens when brevity makes a reply unclear?

Ambiguous clauses use plain Chinese. Confusion, security warnings, and irreversible-action confirmation use full sentences in plain Chinese or the user language.