clash-routes
The clash-routes skill queries active proxy connections via Mihomo API to display which proxy route a specified process uses. It retrieves the Clash Verge configuration, authenticates via Unix socket or HTTP, and presents connections grouped by process name with their associated rules and proxy chains. Use this to verify routing paths for applications like Claude or Chrome through active proxy subscriptions.
git clone --depth 1 https://github.com/majiayu000/spellbook /tmp/clash-routes && cp -r /tmp/clash-routes/skills/clash-routes ~/.claude/skills/clash-routesSKILL.md
# Clash 线路查看工具
查看本机当前活跃连接,确认指定进程命中的规则、策略组与真实出口。不接受 SSH 参数;远程机器应先通过 Tailscale SSH 登录,再在目标机器执行同一只读流程。
用户传入的参数:$ARGUMENTS。没有参数时列出所有活跃连接。诊断 Gemini、ChatGPT 或浏览器流量时,先不加过滤获取实际 `metadata.process`,再使用观察到的进程名;不要假定它们属于 `claude`。
## 获取凭证
读取 Clash Verge 配置,但不要打印 secret:
```bash
SECRET=$(grep '^secret:' "$HOME/Library/Application Support/io.github.clash-verge-rev.clash-verge-rev/clash-verge.yaml" 2>/dev/null | awk '{print $2}')
[ -z "$SECRET" ] && SECRET=$(grep '^secret:' "$HOME/.config/clash/config.yaml" 2>/dev/null | awk '{print $2}')
[ -n "$SECRET" ] && echo "API secret: configured" || echo "API secret: not configured"
```
## 查询与回退
请求成功才采用该 endpoint。`/tmp` socket 存在但失效时继续尝试 `/var/tmp`,最后尝试配置的 HTTP controller。所有 endpoint 都失败时明确报错,不返回空数据。
```bash
request_connections() {
for socket_path in \
/tmp/verge/verge-mihomo.sock \
/var/tmp/verge/verge-mihomo.sock
do
[ -S "$socket_path" ] || continue
if curl --fail --silent --show-error \
--unix-socket "$socket_path" \
"http://localhost/connections" \
-H "Authorization: Bearer $SECRET"
then
return 0
fi
done
controller=$(grep '^external-controller:' "$HOME/Library/Application Support/io.github.clash-verge-rev.clash-verge-rev/clash-verge.yaml" 2>/dev/null | awk '{print $2}' | tr -d "'\"")
[ -n "$controller" ] || controller="127.0.0.1:9090"
curl --fail --silent --show-error \
"http://$controller/connections" \
-H "Authorization: Bearer $SECRET"
}
if ! DATA=$(request_connections); then
echo "无法从 Mihomo Unix socket 或 HTTP controller 获取连接数据" >&2
exit 1
fi
if ! printf '%s' "$DATA" | python3 -c 'import json, sys; value=json.load(sys.stdin); assert isinstance(value.get("connections"), list)' 2>/dev/null; then
echo "Mihomo 返回了无效的 connections JSON" >&2
exit 1
fi
```
## 解析并展示
通过环境变量传递过滤值,避免把用户输入插进 Python 源码:
```bash
printf '%s' "$DATA" | FILTER="$ARGUMENTS" python3 -c '
import json
import os
import sys
from collections import defaultdict
data = json.load(sys.stdin)
process_filter = os.environ.get("FILTER", "").strip().lower()
results = []
for connection in data.get("connections", []):
metadata = connection.get("metadata", {})
process = metadata.get("process", "unknown")
if process_filter and process_filter not in process.lower():
continue
host = metadata.get("host", "") or metadata.get("destinationIP", "")
port = metadata.get("destinationPort", "")
rule = connection.get("rule", "")
payload = connection.get("rulePayload", "")
if payload:
rule += "/" + payload
chains = connection.get("chains", [])
chain_text = " → ".join(reversed(chains)) if chains else "DIRECT"
results.append({
"process": process,
"host": f"{host}:{port}" if port else host,
"rule": rule,
"chain": chain_text,
})
grouped = defaultdict(list)
for result in results:
grouped[result["process"]].append(result)
if not grouped:
target = process_filter or "任何进程"
print(f"未找到 {target} 的活跃连接")
raise SystemExit(0)
for process, connections in sorted(grouped.items()):
print(f"\n进程: {process} ({len(connections)} 个连接)")
route_stats = defaultdict(lambda: {"count": 0, "hosts": set()})
for connection in connections:
key = "{} → {}".format(connection["rule"], connection["chain"])
route_stats[key]["count"] += 1
route_stats[key]["hosts"].add(connection["host"])
for route, info in sorted(route_stats.items(), key=lambda item: -item[1]["count"]):
hosts = sorted(info["hosts"])
shown = ", ".join(hosts[:5])
suffix = f" ... (+{len(hosts) - 5})" if len(hosts) > 5 else ""
print(f" 线路: {route}")
print(" 连接数: {}".format(info["count"]))
print(f" 目标: {shown}{suffix}")
'
```
## 解释结果
- `chains[-1]` 是命中的策略组,`chains[0]` 是真实出口;展示时反转为“策略组 → 出口”。
- Claude 常见进程是 `claude` 或 `Claude Helper`;Codex CLI 常见进程是 `codex`。
- Gemini CLI、ChatGPT 桌面端和浏览器的进程名以未过滤连接表为准。
- Host 只有裸 IP 时,检查 `sniffer.parse-pure-ip` 和系统 DNS 是否绕过 Clash,不要直接改 AI 策略组。
- 当前 profile 不是预期 Hub,或出口仍是原始订阅节点时,报告 drift;本 skill 不修改 YAML。
## 完成条件
- 至少一个 Mihomo endpoint 返回合法 connections JSON。
- 输出明确显示进程、规则、策略组和出口,或明确说明目标进程当前没有活跃连接。
- 全程只读且没有打印 API secret。Audit AND optimize a CLAUDE.md / AGENTS.md instruction file — score it against the five high-leverage patterns, flag anti-patterns, then apply approved fixes in place. Use when the user says 优化 CLAUDE.md / 优化 AGENTS.md / optimize my agent doc / 帮我改 claudemd, or after an audit when they want the fixes applied (not just reported).
Generate or update repository-specific AGENTS.md instruction files from real repo evidence. Use when asked to create, design, scaffold, split, or improve root or scoped AGENTS.md files for Codex/Claude/agent workflows, especially when a repo needs directory-specific rules, validation commands, generated-file boundaries, or a short agent onboarding router.
REST/GraphQL/gRPC API design best practices. Use when designing APIs, defining contracts, handling versioning. Covers OpenAPI 3.2, GraphQL Federation, gRPC streaming.
Mobile app UI design expert for iOS and Android. Use when designing app interfaces, creating design systems, ensuring accessibility, or following platform guidelines. Covers Material Design 3, Human Interface Guidelines, color theory, typography, and 2025 trends.
End-to-end app feature inventory and user-story testing workflow with a canonical tracker. Use when the user asks to audit every feature, derive expected behavior from code, test user journeys, or explicitly fix and retest documented UX or logistical defects.
Design architecture foundations before implementation. Use when asked to design or refactor architecture, choose Rust/Go crate, package, module, runtime, workflow, or service boundaries, compare mature project architecture, prevent stacked one-off PRs, audit migration debt in existing repos, or produce a spec before coding.
Evidence-driven architecture research for understanding real systems and making technical decisions. Use when doing architecture landscape studies, source-backed system archaeology, build-vs-buy or adopt/adapt/build decisions, open-source and commercial comparisons, revisiting an earlier architecture choice, or handling requests such as 架构调研, 架构选型, 竞品架构, 技术尽调, 同类方案, 开源替代, how is X built, and what should we learn from X. Do not use for small mechanical changes, market-only discovery, or detailed design after the technology direction is already fixed.
Ask Grok or Gemini via opencli (browser-session based, no API cost). Use when the user says "问 grok", "问 gemini", "ask grok", "ask gemini", "用 grok 查", "grok 怎么看", "让 gemini 分析", or wants a second opinion from Grok/Gemini without paying API tokens. Routes to opencli which drives the already-logged-in Chrome session.