-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathtest.js
More file actions
206 lines (189 loc) · 10.4 KB
/
Copy pathtest.js
File metadata and controls
206 lines (189 loc) · 10.4 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
'use strict'
/**
* server.js 的进程内测试：直接调 handleMessage，不经 stdio 管道，
* 避免沙箱下子进程 stdio 管道的 EPERM 边界，同时逻辑覆盖面完全一致。
* 本地 HTTP 服务器测试全部管线不依赖外网；外网用例失败只警告不判败。
*/
const http = require('http')
const { handleMessage } = require('./server.js')
function roundtrip(msg, timeoutMs = 45000) {
return new Promise((resolve) => {
const timer = setTimeout(() => resolve({ timeout: true }), timeoutMs)
handleMessage(JSON.stringify(msg), (resp) => {
clearTimeout(timer)
resolve(resp)
})
})
}
const results = []
function check(name, ok, detail = '') {
results.push({ name, ok })
console.log(`${ok ? 'PASS' : 'FAIL'} ${name}${detail ? ' — ' + detail : ''}`)
}
function call(args, id = 100) {
return roundtrip({ jsonrpc: '2.0', id, method: 'tools/call', params: { name: 'fetch', arguments: args } })
}
function textOf(r) {
return (r.result && r.result.content && r.result.content[0] && r.result.content[0].text) || ''
}
(async () => {
// ===== 协议层 =====
let r = await roundtrip({ jsonrpc: '2.0', id: 1, method: 'initialize', params: { protocolVersion: '2025-06-18', capabilities: {}, clientInfo: { name: 't', version: '0' } } })
check('initialize 回显协议版本', r.result?.protocolVersion === '2025-06-18' && r.result?.serverInfo?.name === 'dsh-webfetch')
r = await roundtrip({ jsonrpc: '2.0', id: 2, method: 'initialize', params: { protocolVersion: '1999-01-01', capabilities: {}, clientInfo: {} } })
check('initialize 未知版本回落', r.result?.protocolVersion === '2025-03-26')
r = await roundtrip({ jsonrpc: '2.0', id: 3, method: 'tools/list', params: {} })
const schema = r.result?.tools?.[0]?.inputSchema?.properties || {}
check('tools/list 只有 fetch 且含新参数', r.result?.tools?.map((t) => t.name).join() === 'fetch' && !!schema.startIndex && schema.format.enum.join() === 'text,markdown,html,json')
let notified = false
handleMessage(JSON.stringify({ jsonrpc: '2.0', method: 'notifications/initialized' }), () => { notified = true })
check('notification 不回复', notified === false)
handleMessage('{bad json', (resp) => { check('非法 JSON 行返回 -32700', resp.error?.code === -32700 && resp.id === null) })
r = await roundtrip({ jsonrpc: '2.0', id: 6, method: 'resources/list' })
check('未知方法 -32601', r.error?.code === -32601)
// ===== 参数校验与 SSRF =====
r = await call({})
check('缺 url 报错', r.result?.isError === true)
r = await call({ url: 'file:///C:/Windows/win.ini' })
check('file:// 协议被拒', r.result?.isError === true && textOf(r).includes('仅支持 http/https'))
r = await call({ url: 'https://x.test/', format: 'yaml' })
check('非法 format 报错', r.result?.isError === true)
r = await call({ url: 'http://169.254.169.254/latest/meta-data/' })
check('SSRF：链路本地地址被拒', r.result?.isError === true && textOf(r).includes('安全考虑'))
r = await call({ url: 'http://metadata.google.internal/' })
check('SSRF：云元数据被拒', r.result?.isError === true && textOf(r).includes('安全考虑'))
// ===== 本地 HTTP 服务器 =====
const seen = { ua: {}, accept: {} }
// 真实页面结构：标题/导语/链接/图片都在 article 内部，导航噪声在外部
const bigArticle =
'<article>' +
'<h1>标题一</h1>' +
'<p>第一段 <a href="/rel">链接A</a> <img src="/i.png" alt="图甲"> 内容。</p>' +
'<p><a href=/unquoted>裸链接</a> <a href="/logo"><img src="/logo.png" alt="站点logo"></a></p>' +
'<p>第二段 & 实体 😀</p>' +
'<p>正文内容甲乙丙丁。</p>'.repeat(400) +
'</article>'
const navNoise = '<nav>' + '导航噪声一二三。'.repeat(3) + '</nav>'
const sample = `<!DOCTYPE html><html><head><title>测试页</title><style>.x{color:red}</style></head><body>
${navNoise}<script>var evil=1;</script>
${bigArticle}</body></html>`
let cfHits = 0
const longBody = '长文内容。'.repeat(1200) // 约 6000 字符
const server = http.createServer((req, res) => {
seen.ua[req.url] = req.headers['user-agent'] || ''
seen.accept[req.url] = req.headers.accept || ''
if (req.url === '/robots.txt') {
res.setHeader('content-type', 'text/plain; charset=utf-8')
res.end('User-agent: *\nDisallow: /private\n')
} else if (req.url.startsWith('/private/')) {
res.setHeader('content-type', 'text/html; charset=utf-8')
res.end('<html><body>秘密内容不应被返回</body></html>')
} else if (req.url === '/cf') {
cfHits++
if (cfHits === 1) {
res.statusCode = 403
res.setHeader('cf-mitigated', 'challenge')
res.setHeader('content-type', 'text/html; charset=utf-8')
res.end('<html><body>Just a moment...</body></html>')
} else {
res.setHeader('content-type', 'text/html; charset=utf-8')
res.end('<html><body>挑战通过，正常内容</body></html>')
}
} else if (req.url === '/json') {
res.setHeader('content-type', 'application/json; charset=utf-8')
res.end(JSON.stringify({ ok: true, items: [1, 2, 3] }))
} else if (req.url === '/long') {
res.setHeader('content-type', 'text/html; charset=utf-8')
res.end(`<html><head><title>长文</title></head><body>${longBody}</body></html>`)
} else if (req.url === '/echo') {
res.setHeader('content-type', 'application/json; charset=utf-8')
res.end(JSON.stringify({ accept: req.headers.accept, ua: req.headers['user-agent'] }))
} else {
res.setHeader('content-type', 'text/html; charset=utf-8')
res.end(sample)
}
})
await new Promise((resolve) => server.listen(0, '127.0.0.1', resolve))
const base = `http://127.0.0.1:${server.address().port}`
// ===== HTML 清洗与正文优先 =====
r = await call({ url: base + '/' })
let t = textOf(r)
check('清洗：无 script/style', !t.includes('evil') && !t.includes('.x{'))
check('清洗：保留标题与正文', t.includes('标题: 测试页') && t.includes('标题一') && t.includes('第一段'))
check('清洗：实体解码', t.includes('& 实体 😀'))
check('正文优先：main/article 命中后无导航噪声', t.includes('正文内容甲乙丙丁') && !t.includes('导航噪声'))
// ===== markdown：链接绝对化 + 图片 alt =====
r = await call({ url: base + '/', format: 'markdown' })
t = textOf(r)
check('markdown：相对链接绝对化', t.includes(`[链接A](${base}/rel)`), t.split('\n').find((l) => l.includes('链接A')) || '')
check('markdown：图片转 ', t.includes(``))
check('markdown：无引号 href 也能绝对化', t.includes(`[裸链接](${base}/unquoted)`))
check('markdown：图片型链接用 alt 当文本', t.includes(`[站点logo](${base}/logo)`))
check('markdown：标题保留', t.includes('# 标题一'))
// ===== html 原样 / JSON / 截断 / 分块 =====
r = await call({ url: base + '/', format: 'html', maxChars: 3000 })
t = textOf(r)
check('html 格式原样返回', t.includes('<nav') || t.includes('<article'))
r = await call({ url: base + '/json' })
t = textOf(r)
check('JSON 自动格式化', !r.result?.isError && t.includes('"items": [\n 1,'))
r = await call({ url: base + '/long', maxChars: 1000 })
t = textOf(r)
check('分块第一片：有后文标记', t.includes('后文未显示') && t.includes('startIndex=1000'))
const firstLen = t.length
r = await call({ url: base + '/long', maxChars: 1000, startIndex: 1000 })
t = textOf(r)
check('分块第二片：有前文标记且长度相近', t.includes('自第 1000 字符起') && Math.abs(t.length - firstLen) < 120)
r = await call({ url: base + '/long', maxChars: 1000, startIndex: 10000 })
t = textOf(r)
check('startIndex 超界：明确越界提示', t.includes('已超出内容范围') && !t.includes('后文未显示'))
// ===== robots.txt =====
r = await call({ url: base + '/private/page' })
check('robots：Disallow 路径被拒', r.result?.isError === true && textOf(r).includes('robots.txt'))
r = await call({ url: base + '/public/page' })
check('robots：允许路径放行', !r.result?.isError)
// 空 Disallow（RFC 9309 语义=全允许）与 Disallow: /（全封）各用一个独立 origin，避开缓存
const mk = (robotsText, bodyText) =>
new Promise((resolve) => {
const s = http.createServer((req, res) => {
if (req.url === '/robots.txt') {
res.setHeader('content-type', 'text/plain; charset=utf-8')
res.end(robotsText)
} else {
res.setHeader('content-type', 'text/html; charset=utf-8')
res.end(`<html><body>${bodyText}</body></html>`)
}
})
s.listen(0, '127.0.0.1', () => resolve({ s, base: `http://127.0.0.1:${s.address().port}` }))
})
const emptyDisallow = await mk('User-agent: *\nDisallow:\n', '应该放行')
r = await call({ url: emptyDisallow.base + '/page' })
check('robots：空 Disallow 视为全允许', !r.result?.isError && textOf(r).includes('应该放行'))
emptyDisallow.s.close()
const slashDisallow = await mk('User-agent: *\nDisallow: /\n', '不该放行')
r = await call({ url: slashDisallow.base + '/page' })
check('robots：Disallow / 全封', r.result?.isError === true && textOf(r).includes('robots.txt'))
slashDisallow.s.close()
// ===== Cloudflare 挑战重试 =====
r = await call({ url: base + '/cf' })
t = textOf(r)
check('CF 挑战：自动换 UA 重试成功', cfHits === 2 && t.includes('挑战通过') && !r.result?.isError)
check('CF 重试使用了不同 UA', seen.ua['/cf'] === 'dsh-webfetch/1.1 (+https://harness.deepseek.com)')
// ===== Accept 头按格式 =====
await call({ url: base + '/echo', format: 'markdown' })
check('Accept：markdown 格式请求 markdown 优先', (seen.accept['/echo'] || '').startsWith('text/markdown'))
server.close()
// ===== 外网探活（失败仅警告）=====
try {
r = await call({ url: 'https://example.com/', timeoutMs: 15000 })
console.log(`WARN 外网 example.com：${r.result?.isError ? '失败(' + textOf(r).slice(0, 80) + ')' : '成功'}`)
} catch (e) {
console.log(`WARN 外网用例异常：${e.message}`)
}
const failed = results.filter((x) => !x.ok)
console.log(`\n${results.length - failed.length}/${results.length} 通过`)
process.exit(failed.length ? 1 : 0)
})().catch((e) => {
console.error('测试运行异常：', e)
process.exit(1)
})