breath57/dingtalk-skills / .agents/skills/dingtalk-ai-web-search
dingtalk-ai-web-search
Verified网页搜索与实时信息检索。需要查找最新资讯、实时数据、技术文档、热点事件,或用户说"搜一下"、"帮我查"、"查资料"、"web search"等时使用。基于小宿AI智能搜索,支持关键词搜索、时间范围过滤(一天/一周/一月/一年)、自定义结果数量、JSON 输出。
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 3.6 KB |
| scripts/search.shexecutable | script | 11.9 KB |
scripts/search.sh
View on GitHub#!/usr/bin/env bash
# web_search_v3/scripts/search.sh
# 网页搜索脚本(小宿智能搜索 MCP 网关)
# 依赖:curl(+ bash/grep/sed 等标准系统工具,无需任何 npm install / pip install)
#
# 用法:
# bash search.sh -q "搜索词" [选项]
#
# 选项:
# -q, --query 搜索词(正常搜索时必填)
# -c, --config MCP JSON 配置字符串(配合 --save 使用)
# -n, --count 返回条数(默认: 5,最大 50)
# -f, --freshness noLimit|oneDay|oneWeek|oneMonth|oneYear(默认: noLimit)
# --json 以 JSON 格式输出
# --ping 仅测试连通性,列出可用工具后退出
# --save 对 -c 指定的配置做连通性测试,通过后永久写入本脚本
set -euo pipefail
SCRIPT_PATH="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd -P)/$(basename "${BASH_SOURCE[0]}")"
# JSON 解析工具优先级:jq > python3/python > bare grep
JQ=""
command -v jq &>/dev/null && JQ=jq
PYTHON=""
if command -v python3 &>/dev/null; then
PYTHON=python3
elif command -v python &>/dev/null; then
PYTHON=python
fi
# 配置文件路径(URL 存储在脚本旁边,不自修改脚本本身)
CONFIG_FILE="$(dirname "$SCRIPT_PATH")/.mcp_url"
DEFAULT_MCP_URL=""
[[ -f "$CONFIG_FILE" ]] && DEFAULT_MCP_URL="$(cat "$CONFIG_FILE")" || true
# ── 工具函数 ──────────────────────────────────────────────────────────────────
die() { echo "错误:$*" >&2; exit 1; }
# JSON 字符串转义(bash 内建,无外部依赖)
json_str() {
local s="$1"
s="${s//\\/\\\\}" # \ → \\
s="${s//\"/\\\"}" # " → \"
s="${s//$'\n'/\\n}" # newline
s="${s//$'\r'/\\r}" # CR
s="${s//$'\t'/\\t}" # tab
printf '"%s"' "$s"
}
# 保存 URL 到配置文件(不修改脚本本身,彻底避免 bash 缓冲读取问题)
save_url() {
local url="$1"
printf '%s' "$url" > "$CONFIG_FILE" || die "写入配置文件失败,请检查文件权限($CONFIG_FILE)"
}
# 解析 MCP JSON 配置,提取 url 字段
parse_mcp_url() {
local cfg="$1"
if [[ -n "$JQ" ]]; then
echo "$cfg" | jq -re '[.mcpServers[].url] | first' 2>/dev/null \
|| die "配置中未找到有效的 url 字段,请确认 JSON 格式正确"
elif [[ -n "$PYTHON" ]]; then
$PYTHON -c "
import sys, json
try:
cfg = json.loads(sys.argv[1])
for srv in cfg.get('mcpServers', {}).values():
if srv.get('url'):
print(srv['url'])
sys.exit(0)
sys.exit(1)
except Exception as e:
print(str(e), file=sys.stderr)
sys.exit(1)
" "$cfg" || die "配置中未找到有效的 url 字段,请确认 JSON 格式正确"
else
echo "$cfg" \
| grep -o '"url"[[:space:]]*:[[:space:]]*"[^"]*"' \
| head -1 \
| sed 's/"url"[[:space:]]*:[[:space:]]*"//;s/"$//' \
|| die "配置中未找到有效的 url 字段"
fi
}
# ── MCP Streamable HTTP 请求 ──────────────────────────────────────────────────
# 发送 MCP POST,返回响应 JSON(自动处理 SSE)
# 用法: mcp_post <url> <body> [session_id]
mcp_post() {
local url="$1" body="$2" sid="${3:-}"
local hdr_file ct resp
hdr_file=$(mktemp) || die "无法创建临时文件"
local -a cmd=(
curl -s -S -m 30 --connect-timeout 10
-X POST
-H 'Content-Type: application/json'
-H 'Accept: application/json, text/event-stream'
-D "$hdr_file"
--data-raw "$body"
)
[[ -n "$sid" ]] && cmd+=(-H "Mcp-Session-Id: $sid")
cmd+=("$url")
if ! resp=$("${cmd[@]}" 2>&1); then
rm -f "$hdr_file"
die "$resp"
fi
ct=$(grep -i '^content-type:' "$hdr_file" | tail -1 \
| tr -d '\r\n' | sed 's/^[Cc]ontent-[Tt]ype:[[:space:]]*//' || true)
rm -f "$hdr_file"
if [[ "$ct" == *event-stream* ]]; then
# 从 SSE 流中取最后一条含 result / error 的 data 行
resp=$(printf '%s\n' "$resp" \
| grep '^data: ' \
| grep -E '"result"|"error"' \
| tail -1 \
| cut -c7-)
fi
printf '%s\n' "$resp"
}
# 初始化 MCP 会话,输出 Session-ID(可为空)
mcp_init() {
local url="$1"
local hdr_file
hdr_file=$(mktemp) || die "无法创建临时文件"
curl -s -S -m 30 --connect-timeout 10 \
-X POST \
-H 'Content-Type: application/json' \
-H 'Accept: application/json, text/event-stream' \
-D "$hdr_file" \
--data-raw '{"jsonrpc":"2.0","id":1,"method":"initialize","params":{"protocolVersion":"2024-11-05","capabilities":{},"clientInfo":{"name":"web-search","version":"1.0.0"}}}' \
"$url" > /dev/null 2>&1 || true
grep -i '^mcp-session-id:' "$hdr_file" \
| tr -d '\r\n' \
| sed 's/^[Mm][Cc][Pp]-[Ss]ession-[Ii][Dd]:[[:space:]]*//' \
|| true
rm -f "$hdr_file"
}
# 提取 JSON-RPC 错误消息(有错误时输出,否则空)
rpc_error() {
echo "$1" | grep -o '"message":"[^"]*"' | head -1 \
| sed 's/"message":"//;s/"$//' || true
}
# ── 命令实现 ──────────────────────────────────────────────────────────────────
cmd_ping() {
local url="$1"
local sid resp tools
sid=$(mcp_init "$url")
mcp_post "$url" \
'{"jsonrpc":"2.0","method":"notifications/initialized","params":{}}' \
"$sid" > /dev/null 2>&1 || true
resp=$(mcp_post "$url" '{"jsonrpc":"2.0","id":2,"method":"tools/list","params":{}}' "$sid")
if echo "$resp" | grep -q '"error"'; then
local msg; msg=$(rpc_error "$resp")
die "连接失败:${msg:-未知错误}"
fi
if [[ -n "$JQ" ]]; then
tools=$(echo "$resp" | jq -r '[.result.tools[].name] | join(", ")' 2>/dev/null || echo "(无法解析)")
elif [[ -n "$PYTHON" ]]; then
tools=$($PYTHON -c "
import sys, json
d = json.loads(sys.argv[1])
names = [t['name'] for t in d.get('result', {}).get('tools', [])]
print(', '.join(names) if names else '(无工具)')
" "$resp")
else
tools=$(echo "$resp" \
| grep -o '"name":"[^"]*"' \
| sed 's/"name":"//;s/"$//' \
| tr '\n' ',' | sed 's/,$//' || echo "(无法解析)")
fi
echo "✓ 连通成功,可用工具:$tools"
}
cmd_search() {
local url="$1" query="$2" count="${3:-5}" freshness="${4:-noLimit}" as_json="${5:-false}"
# freshness 映射
local fv=""
case "$freshness" in
oneDay) fv="Day" ;;
oneWeek) fv="Week" ;;
oneMonth) fv="Month" ;;
oneYear) fv="Year" ;;
esac
# 构建 tool_arguments JSON
local q_json; q_json=$(json_str "$query")
local tool_args="{\"q\":${q_json},\"count\":${count}"
[[ -n "$fv" ]] && tool_args+=",\"freshness\":\"${fv}\""
tool_args+="}"
local call_body
call_body="{\"jsonrpc\":\"2.0\",\"id\":3,\"method\":\"tools/call\",\"params\":{\"name\":\"web_search\",\"arguments\":${tool_args}}}"
# 初始化会话
local sid
sid=$(mcp_init "$url")
mcp_post "$url" \
'{"jsonrpc":"2.0","method":"notifications/initialized","params":{}}' \
"$sid" > /dev/null 2>&1 || true
# 调用工具
local resp
resp=$(mcp_post "$url" "$call_body" "$sid")
if echo "$resp" | grep -q '"error"'; then
local msg; msg=$(rpc_error "$resp")
die "搜索失败:${msg:-未知错误}"
fi
# 提取 content[0].text
local text
if [[ -n "$JQ" ]]; then
text=$(echo "$resp" | jq -r '.result.content[0].text' 2>/dev/null) || die "无法解析搜索结果"
elif [[ -n "$PYTHON" ]]; then
text=$($PYTHON -c "
import sys, json
d = json.loads(sys.argv[1])
print(d['result']['content'][0]['text'])
" "$resp") || die "无法解析搜索结果"
else
# grep -P 处理 JSON 转义字符串(GNU grep);失败则输出原始响应
# 正则 (\\.|[^"])* 匹配:\"(转义引号)或任意非引号字符
text=$(echo "$resp" \
| grep -oP '"text":"\K(\\.|[^"])*' \
| head -1 \
| sed 's/\\"/"/g;s/\\\\/\\/g' 2>/dev/null) || text=""
fi
# 格式化并输出结果
if [[ -n "$JQ" ]]; then
if $as_json; then
echo "$text" | jq -r '
[ .webPages.value[]? // .result.webPages.value[]? | {
title: (.name // .title // ""),
url: (.url // ""),
snippet: ((.snippet // .summary // "")[:300]),
site: (.siteName // ""),
published: ((.datePublished // "")[:10]),
source: "小宿智能搜索"
}] ' 2>/dev/null || echo "$text"
else
echo "$text" | jq -r '
(.webPages.value // .result.webPages.value // []) | to_entries[] |
"[\(.key + 1)] \(.value.name // .value.title // "")\n URL: \(.value.url // "")\n 摘要: \((.value.snippet // .value.summary // "")[:200])\n"' \
2>/dev/null || echo "$text"
fi
elif [[ -n "$PYTHON" ]]; then
$PYTHON -c "
import sys, json
text = sys.argv[1]
as_json = sys.argv[2] == 'true'
try:
data = json.loads(text)
except Exception:
print(text)
sys.exit(0)
pages = data.get('webPages', {}).get('value', [])
if not pages:
pages = data.get('result', {}).get('webPages', {}).get('value', [])
if not pages:
print(text[:1000] if not as_json else json.dumps({'raw': text[:1000]}, ensure_ascii=False))
sys.exit(0)
if as_json:
out = []
for p in pages:
out.append({
'title': p.get('name') or p.get('title', ''),
'url': p.get('url', ''),
'snippet': (p.get('snippet') or p.get('summary', ''))[:300],
'site': p.get('siteName', ''),
'published': (p.get('datePublished') or '')[:10],
'source': '小宿智能搜索',
})
print(json.dumps(out, ensure_ascii=False, indent=2))
else:
for i, p in enumerate(pages, 1):
title = p.get('name') or p.get('title', '')
url = p.get('url', '')
snippet = (p.get('snippet') or p.get('summary', ''))[:200]
site = p.get('siteName', '')
pub = (p.get('datePublished') or '')[:10]
print(f'[{i}] {title}')
if url: print(f' URL: {url}')
if snippet: print(f' 摘要: {snippet}')
if site: print(f' 来源: {site} (小宿智能搜索)')
if pub: print(f' 发布: {pub}')
print()
" "$text" "$as_json"
else
# 最终兜底:输出原始文本,提示安装 python 或 jq
echo "(提示:安装 python 或 jq 可获得格式化输出,当前输出原始结果)" >&2
echo "$text"
fi
}
# ── 主函数(包裹全部逻辑,bash 定义函数时整体读入内存,避免自写文件时触发 EOF)──
main() {
local ARG_QUERY="" ARG_CONFIG="" ARG_COUNT=5
local ARG_FRESHNESS="noLimit" ARG_JSON=false ARG_PING=false ARG_SAVE=false
while [[ $# -gt 0 ]]; do
case "$1" in
-q|--query) ARG_QUERY="$2"; shift 2 ;;
-c|--config) ARG_CONFIG="$2"; shift 2 ;;
-n|--count) ARG_COUNT="$2"; shift 2 ;;
-f|--freshness) ARG_FRESHNESS="$2"; shift 2 ;;
--json) ARG_JSON=true; shift ;;
--ping) ARG_PING=true; shift ;;
--save) ARG_SAVE=true; shift ;;
*) die "未知参数:$1(使用 -q/--query、--ping、--save 等)" ;;
esac
done
if $ARG_SAVE; then
[[ -z "$ARG_CONFIG" ]] && die "--save 需要配合 -c 传入 MCP JSON 配置"
local MCP_URL
MCP_URL=$(parse_mcp_url "$ARG_CONFIG")
echo "正在检测连通性..."
cmd_ping "$MCP_URL"
save_url "$MCP_URL"
echo "✓ 配置已保存,后续调用无需再传 -c 参数。" >&2
elif $ARG_PING; then
[[ -z "$DEFAULT_MCP_URL" ]] && die "未配置 MCP URL,请先使用 --save 完成配置"
cmd_ping "$DEFAULT_MCP_URL"
else
[[ -z "$ARG_QUERY" ]] && die "请使用 -q 指定搜索词"
[[ -z "$DEFAULT_MCP_URL" ]] && die "未配置 MCP URL,请先使用 --save 完成配置"
cmd_search "$DEFAULT_MCP_URL" "$ARG_QUERY" "$ARG_COUNT" "$ARG_FRESHNESS" "$ARG_JSON"
fi
}
main "$@"
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| add-product-e2e-evalpaperclipai/paperclipAdd or extend a Paperclip full-stack runner E2E workflow, fixture, matcher, or report evidence path for local or Daytona execution. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |
| add-runner-evalpaperclipai/paperclipAdd or extend a Paperclip Runner protocol evaluation definition, roster, assertion, or report fixture with provenance and narrow validation. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |
| diagnose-why-work-stoppedpaperclipai/paperclipCompatibility entry point for failed, stalled, looping, or over-recovered Paperclip work. Use fix-failed-paperclip-run for the full diagnosis, regression, repair, real verification, and green-PR workflow; honor an explicitly requested diagnosis-only scope. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |
| fix-failed-paperclip-runpaperclipai/paperclipDiagnose and fix failed, stuck, slow, or incorrectly blocked Paperclip tasks and runs. Use with a task/run URL or failure report to trace the root cause, prove a regression red then green, check second-order effects, verify the real workflow, and deliver a PR with passing checks. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |