China Mirror Resolver

What it does

When your agent uses it

• Connection refused
• Pip install failed

Example prompts

• “timeout”
• “mirror”
• “registry”
• “/china-mirror-resolver”

Requirements

• Python 3
• A Bash shell
• Docker
• Compatibility (from SKILL.md): Requires terminal/shell access and internet connectivity. Works on Windows (PowerShell), macOS (bash/zsh), and Linux (bash). Designed for Claude Code, StepFun, and similar agentic coding assistants. Adaptable to Cursor (.mdc rules) and Windsurf (.md rules) — see Section 9. Agents without web search capability can still use baseline sources (Section 2).
• Pre-approved tools (allowed-tools): WebSearch, WebFetch, terminal, file-read, file-write

Workflow steps

9 steps, taken from the step headings in SKILL.md.

1. 1Diagnostic Table
2. 2Baseline Sources
3. 3Search Strategy
4. 4Validation
5. 5Configuration
6. 6Restore Official Sources
7. 7Notes
8. 8Known Unstable Sources
9. 9Cross-Agent Adaptation

What it can do on your machine

Read from SKILL.md and the folder at commit e5199b5. It shows what the files ask for, not the result of running them.

• Tool permissions

  Pre-approves these tools, so the agent can use them without asking each time:

  • WebSearch
  • WebFetch
  • terminal
  • file-read
  • file-write

  From allowed-tools in the SKILL.md frontmatter.

• Runs code

  Ships 1 file in scripts/ (Shell), which the agent can run.

  Shell commands in SKILL.md call:

  • npm
  • yarn
  • conda
  • pip
  • pnpm
  • go
  • bash
  • docker
  • git
  • cargo
  • apt
  • yum

  From the folder's file list and the shell code blocks in SKILL.md.

• Network

  Hosts in commands or code, which the agent is likely to contact:

  • mirrors.tuna.tsinghua.edu.cn
  • mirrors.ustc.edu.cn
  • mirrors.aliyun.com
  • repo.huaweicloud.com
  • rsproxy.cn
  • maven.aliyun.com
  • registry.npmjs.org
  • registry.yarnpkg.com
  • pypi.tuna.tsinghua.edu.cn
  • pypi.mirrors.ustc.edu.cn
  • mirrors.cloud.tencent.com
  • registry.npmmirror.com
  • docker.1ms.run
  • docker.xuanyuan.me
  • docker.m.daocloud.io

  From URLs in SKILL.md, links to its own repository left out.

• Credentials

  Names no API keys, tokens, secrets or passwords.

  From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

• Compatibility

  Requires terminal/shell access and internet connectivity. Works on Windows (PowerShell), macOS (bash/zsh), and Linux (bash). Designed for Claude Code, StepFun, and similar agentic coding assistants. Adaptable to Cursor (.mdc rules) and Windsurf (.md rules) — see Section 9. Agents without web search capability can still use baseline sources (Section 2).

  From compatibility in the SKILL.md frontmatter.

Context cost

China Mirror Resolver loads about 4.6k tokens when it runs, and up to ~6.1k if it reads all its reference files. Until then it costs about 118 tokens; SKILL.md has 1,553 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it

~118

When it runs · the whole SKILL.md, loaded when a task matches

~4.6k

With references · SKILL.md plus every file in references/, read only if the agent opens them

~6.1k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

• NoteRuns commands with sudoSKILL.md:357

  r host in `/etc/apt/sources.list`, then `sudo apt update`

• NoteRuns commands with sudoSKILL.md:359

  , replace `baseurl` in repo files, then `sudo yum makecache`

• NoteRuns commands with sudoSKILL.md:383

  re `/etc/apt/sources.list` from backup, `sudo apt update` |

• NoteRuns commands with sudoSKILL.md:384

  estore `/etc/yum.repos.d/` from backup, `sudo yum makecache` |

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); the scripts in this folder are not scanned.

SKILL.md

The full file from LeoYeAI/openclaw-master-skills at commit e5199b5, republished under its MIT licence (© LeoYeAI). 1,553 words, ~4,556 tokens.

name

china-mirror-resolver

description

Self-healing China mirror source resolver. Automatically discovers, validates, and configures domestic mirror sources for pip/npm/yarn/pnpm/conda/Docker/Go/ Rust/Maven/Gradle/Homebrew/apt/yum/GitHub/HuggingFace when downloads are slow or timing out in mainland China. Trigger keywords: "换源", "镜像源", "下载慢", "超时", "timeout", "mirror", "registry", "proxy", "connection refused", "SSL error", "443", "ETIMEDOUT", "CondaHTTPError", "pip install failed".

allowed-tools

WebSearch, WebFetch, terminal, file-read, file-write

compatibility

Requires terminal/shell access and internet connectivity. Works on Windows (PowerShell), macOS (bash/zsh), and Linux (bash). Designed for Claude Code, StepFun, and similar agentic coding assistants. Adaptable to Cursor (.mdc rules) and Windsurf (.md rules) — see Section 9. Agents without web search capability can still use baseline sources (Section 2).

version

2.1.0

license

MIT

homepage

https://github.com/The-Ladder-of-Rrogress/china-mirror-resolver

metadata.author

community

metadata.category

devtools

metadata.region

china-mainland

metadata.last-verified

2026-03-08

User error / request
      │
      ▼
 1. Identify tool type (Section 1)
      │
      ▼
 2. Detect current source config
      │
      ▼
 3. Try ALL baseline sources for that tool (Section 2)
    Validate each with Section 4 thresholds (< 3s = good, 3–10s = acceptable)
      ├─ At least one source < 10s → pick fastest → go to step 6
      └─ All sources > 10s or failed ↓
      ▼
 4. [IF search capable] Search for latest sources (Section 3)
      ├─ Found candidates → go to step 5
      └─ No candidates found → report to user with baseline table as reference
    [IF no search] Report baseline failure to user, suggest manual URL
      │
      ▼
 5. Validate ALL candidates (Section 4), pick fastest passing one
      ├─ At least one passes → go to step 6
      └─ All fail → report failure to user, suggest checking network/proxy
      │
      ▼
 6. Backup original config → write new config → verify with test command
      │
      ▼
 7. Confirm result to user (show old source → new source, response time)

# Linux / macOS
start=$(date +%s%N 2>/dev/null || perl -MTime::HiRes=time -e 'printf "%d\n",time()*1e9')
code=$(curl -o /dev/null -s -w "%{http_code}" --max-time 10 "<URL>")
end=$(date +%s%N 2>/dev/null || perl -MTime::HiRes=time -e 'printf "%d\n",time()*1e9')
echo "HTTP $code | $(( (end - start) / 1000000 ))ms"

# Windows PowerShell
$sw = [System.Diagnostics.Stopwatch]::StartNew()
try { $r = Invoke-WebRequest -Uri "<URL>" -TimeoutSec 10 -UseBasicParsing -ErrorAction Stop
  $sw.Stop(); "OK: HTTP $($r.StatusCode) | $($sw.ElapsedMilliseconds)ms"
} catch { $sw.Stop(); "FAIL: $_" }

bash scripts/validate.sh              # test all, colored output
bash scripts/validate.sh pip           # test pip only
bash scripts/validate.sh all --json    # machine-readable JSON output

---
description: Resolve slow package downloads in China by switching to domestic mirrors. Triggers on: timeout, mirror, 换源, 镜像源, 下载慢
globs:
alwaysApply: false
---

[Paste the content of this SKILL.md here, excluding the YAML frontmatter]

---
trigger: auto
---

[Paste the content of this SKILL.md here, excluding the YAML frontmatter]

© LeoYeAI, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 6 other files (scripts, references) in skills/china-mirror-resolver of LeoYeAI/openclaw-master-skills.

• SKILL.md
• CHANGELOG.md
• CONTRIBUTING.md
• README.md
• _meta.json
• references/config-templates.md
• scripts/validate.sh

Open the folder on GitHubat commit e5199b5