exchange-owa-attack
Exchange/OWA NTLM AD leak, spray attack when mail subdomain.
Install / Use
npx skills add uphiago/recon-skills --skill exchange-owa-attackInstalls into whichever agent you are using.
SKILL.md
Installable skill definition
Quality Score
Category
Development & EngineeringSupported Platforms
Our assessment of exchange-owa-attack
exchange-owa-attack scores 89/100 on our quality scale, 1653rd of 4,700 Development & Engineering skills we index (top 36%).
Its SKILL.md is 10 KB long, well organised into 32 sections with 5 code examples: a thorough specification that gives an agent plenty to work with.
With 1,280 GitHub stars, it is one of the more widely adopted skills in the catalogue.
Maintenance, license and trust
- The repository was last updated about 3 months ago, so exchange-owa-attack is actively maintained.
- Our last check on 2026-10-31 found the source still online.
- It is released under the MIT license, a permissive license that allows use, modification and commercial use with attribution.
- Its trust signals score 100/100, with no cautions. These come from repository metadata, not a code audit — read the skill file before letting an agent act on it.
Safety scan
ReviewOur scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review judged it risky (a second model did not confirm it as malicious): The skill instructs the agent to capture and decode NTLM Type-2 challenges from OWA to leak Active Directory domain, NetBIOS and computer names.
AI review: risky
- The skill instructs the agent to capture and decode NTLM Type-2 challenges from OWA to leak Active Directory domain, NetBIOS and computer names.
- It includes a 'Password Spray Surface Assessment' that makes the agent send multiple invalid authentication requests to /owa/auth.owa, a credential-guessing/lockout attack step.
- It directs probing of high-value admin/API endpoints (/ecp/, /ews/, /autodiscover/, /mapi/) and ADFS/Office 365 recon to gather intelligence for further compromise.
AI review by kimi-k2.7-code with a second opinion from glm-5.3 on 2026-10-02. Automated pattern scan on 2026-10-01. It catches known dangerous patterns, not every risk — read a skill before letting an agent act on it.
exchange-owa-attack compared with similar skills
All 4 of these similar skills score higher than exchange-owa-attack; compare them before choosing.
| Skill | Score | Stars | Updated | Format |
|---|---|---|---|---|
| exchange-owa-attack (this skill)by uphiago | 89 | 1.3k | 3mo ago | SKILL.md |
| taste-skillby Leonxlnx | 100 | 116.8k | 8d ago | CLAUDE.md |
| Agent-Reachby Panniantong | 100 | 112.3k | 9d ago | CLAUDE.md |
| headroomby headroomlabs-ai | 100 | 79.8k | today | CLAUDE.md |
| ai-job-searchby ai-job-search-dev | 100 | 72.1k | today | CLAUDE.md |
Frequently asked questions
- How do I install exchange-owa-attack?
- Run
npx skills add uphiago/recon-skills --skill exchange-owa-attack. The install tabs above show the steps for each supported agent. - Which AI agents does exchange-owa-attack work with?
- It is written for Universal, as a SKILL.md file. Other agents that read the same format can often use it too.
- Is exchange-owa-attack safe to use?
- Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review judged it risky (a second model did not confirm it as malicious): The skill instructs the agent to capture and decode NTLM Type-2 challenges from OWA to leak Active Directory domain, NetBIOS and computer names. It is MIT-licensed and scores 100/100 on trust signals. Skills are instructions an agent will follow, so read the file before installing it and do not approve commands you do not understand.
- Is exchange-owa-attack still maintained?
- The repository was last updated about 3 months ago, so exchange-owa-attack is actively maintained.
Skill content
View source on GitHubname: exchange-owa-attack description: Exchange/OWA NTLM AD leak, spray attack when mail subdomain. version: 1.1.0 revision_date: 2026-07-25 license: MIT platforms: [linux] compatibility: Requires curl, nmap, python3, masscan, subfinder, httpx, nuclei tags: [recon, exchange, OWA, NTLM, ActiveDirectory, password-spray] category: recon related_skills:
- port-service-discovery
- zimbra-attack
- subdomain-enumeration
Exchange/OWA Attack Skill
Exchange Outlook Web Access reconnaissance covering endpoint mapping, NTLM Type-2 metadata, authentication controls, and version evidence. Password or lockout testing requires explicit authorization and approved identities.
When to Use
- Target has
owa.,mail.,webmail.,exchange., orautodiscover.subdomains. - crt.sh reveals Exchange-related SAN names (
mail.domain.com,autodiscover.domain.com). - Port 443 returns NTLM
WWW-Authenticate: NegotiateorWWW-Authenticate: NTLM. - After
subdomain-enumerationdiscovers mail-related hosts. - After
port-service-discoveryfinds HTTPS on port 443 with Exchange fingerprints.
Prerequisites
terminalwith curl, python3.- Target Exchange/OWA URL.
- For password spray: list of usernames (from recon) and password candidates.
How to Run
# Quick Exchange detection
curl --max-time 30 --connect-timeout 10 -skI "https://TARGET/owa/" | grep -iE "x-owa-version|x-feserver|exchange|microsoft"
# NTLM challenge capture (AD domain leak)
curl --max-time 30 --connect-timeout 10 -skI "https://TARGET/owa/" -H "Authorization: Negotiate TlRMTVNTUAABAAAAB4IIogAAAAAAAAAAAAAAAAAAAAAGAbEdAAAADw==" | grep -i "www-authenticate"
Quick Reference
| Technique | What It Reveals | Severity |
|-----------|---------------|---------|
| NTLM Type-2 decode | AD domain, NetBIOS name, computer name, AD timestamp | High |
| OWA version header | Exchange version, CU level, patch status | Medium |
| /owa/auth/logon.aspx | Login page, brute force surface | Medium |
| /ecp/ | Exchange Control Panel (admin) | High |
| /ews/ | Exchange Web Services (SOAP API) | Medium |
| /autodiscover/ | Autodiscover configuration | Medium |
| /mapi/ | MAPI over HTTP | Low |
| /Microsoft-Server-ActiveSync | Mobile device sync | Medium |
| /rpc/ | Outlook Anywhere (RPC over HTTP) | Low |
Procedure
Phase 1 — Exchange Detection & Fingerprinting
TARGET="$1"
OUTDIR="$OUTDIR/exchange"
mkdir -p "$OUTDIR"
echo "[*] Exchange detection on $TARGET"
# OWA probe
OWA_RESP=$(curl -skI --max-time 10 --connect-timeout 10 "https://$TARGET/owa/" 2>/dev/null)
echo "$OWA_RESP" > "$OUTDIR/owa_headers.txt"
# Version extraction
X_OWA=$(echo "$OWA_RESP" | grep -i "x-owa-version" | sed 's/.*: //')
X_FE=$(echo "$OWA_RESP" | grep -i "x-feserver" | sed 's/.*: //')
if [[ -n "$X_OWA" ]]; then
echo "[+] Exchange confirmed — OWA Version: $X_OWA"
echo " Frontend server: ${X_FE:-unknown}"
# Map version to CU
# 15.1.x = Exchange 2016, 15.2.x = Exchange 2019
MAJOR=$(echo "$X_OWA" | cut -d. -f1-2)
if [[ "$MAJOR" == "15.1" ]]; then
echo " Product: Exchange 2016"
elif [[ "$MAJOR" == "15.2" ]]; then
echo " Product: Exchange 2019"
fi
else
echo "[-] No OWA version header — may not be Exchange"
fi
# Key endpoints probe
declare -A EX_ENDPOINTS
EX_ENDPOINTS["/owa/auth/logon.aspx"]="Login page"
EX_ENDPOINTS["/ecp/"]="Exchange Control Panel (admin)"
EX_ENDPOINTS["/ews/exchange.asmx"]="Exchange Web Services (SOAP)"
EX_ENDPOINTS["/autodiscover/autodiscover.xml"]="Autodiscover"
EX_ENDPOINTS["/mapi/emsmdb/"]="MAPI over HTTP"
EX_ENDPOINTS["/Microsoft-Server-ActiveSync/"]="ActiveSync"
EX_ENDPOINTS["/rpc/rpcproxy.dll"]="Outlook Anywhere"
EX_ENDPOINTS["/owa/healthcheck.htm"]="Health check"
echo ""
echo "[*] Endpoint probe:"
for ep in "${!EX_ENDPOINTS[@]}"; do
code=$(curl -sk -o /dev/null -w "%{http_code}" --max-time 5 --connect-timeout 5 "https://$TARGET$ep")
[[ "$code" == "200" ]] && echo " [OPEN] $ep — ${EX_ENDPOINTS[$ep]}"
[[ "$code" == "302" ]] && echo " [REDIR] $ep — ${EX_ENDPOINTS[$ep]}"
[[ "$code" == "401" ]] && echo " [AUTH] $ep — ${EX_ENDPOINTS[$ep]}"
done
Phase 2 — NTLM Type-2 Challenge Capture & Decode
TARGET="$1"
echo "[*] NTLM challenge capture from $TARGET"
# Send NTLM Type-1 (Negotiate) message via Authorization header
NTLM_RESP=$(curl -skI --max-time 10 --connect-timeout 10 "https://$TARGET/owa/" \
-H "Authorization: Negotiate TlRMTVNTUAABAAAAB4IIogAAAAAAAAAAAAAAAAAAAAAGAbEdAAAADw==" 2>/dev/null)
WWW_AUTH=$(echo "$NTLM_RESP" | grep -i "www-authenticate: negotiate" | sed 's/.*negotiate //i' | tr -d '\r\n ')
if [[ -n "$WWW_AUTH" ]]; then
echo "[+] NTLM Type-2 challenge received!"
echo " Raw: ${WWW_AUTH:0:80}..."
# Decode with Python (extract AV_PAIRS structure)
echo "$WWW_AUTH" | python3 -c "
import base64, struct, sys
data = base64.b64decode(sys.stdin.read().strip())
# NTLM Type-2 message structure:
# Offset 12: Target Name
# Offset 16: Negotiate Flags
# Offset 20: Server Challenge
# Offset 28: Reserved
# Offset 32: Target Info (AV_PAIRS)
# Parse Target Info
if len(data) > 40:
target_info_offset = struct.unpack_from('<I', data, 40)[0]
target_info_len = struct.unpack_from('<I', data, 44)[0]
av_pairs = data[target_info_offset:target_info_offset + target_info_len]
print()
print('=== NTLM Type-2 Decoded ===')
pos = 0
while pos < len(av_pairs) - 4:
av_type = struct.unpack_from('<H', av_pairs, pos)[0]
av_len = struct.unpack_from('<H', av_pairs, pos + 2)[0]
av_value = av_pairs[pos + 4:pos + 4 + av_len]
# AV_PAIR types
types = {
1: 'NetBIOS Computer Name',
2: 'NetBIOS Domain Name',
3: 'DNS Computer Name',
4: 'DNS Domain Name',
5: 'DNS Tree Name',
6: 'Product Version',
7: 'Timestamp',
}
label = types.get(av_type, f'Unknown({av_type})')
if av_type in (1, 2, 3, 4, 5):
value = av_value.decode('utf-16-le', errors='replace')
print(f' {label}: {value}')
elif av_type == 7:
ts = struct.unpack_from('<Q', av_value)[0]
from datetime import datetime, timezone
dt = datetime.fromtimestamp(ts / 10000000 - 11644473600, tz=timezone.utc)
print(f' Timestamp: {dt}')
else:
print(f' {label}: {av_value.hex()}')
pos += 4 + av_len
else:
print(' No AV_PAIRS in response')
"
fi
Phase 3 — Password Spray Surface Assessment
TARGET="$1"
echo "[*] Password spray surface assessment"
# Check for account lockout by testing rapid logins with invalid password
echo "[*] Rate limiting test (5 rapid attempts with wrong password)..."
for i in $(seq 1 5); do
code=$(curl -sk -o /dev/null -w "%{http_code}" --max-time 10 --connect-timeout 10 \
-X POST "https://$TARGET/owa/auth.owa" \
-d "destination=https://$TARGET/owa/&username=testuser$i@domain.com&password=WrongPass123!" 2>/dev/null)
echo " Attempt $i: HTTP $code"
done
# Check if Basic Auth is enabled (rare post-2022, but exists)
BASIC_AUTH=$(curl -skI --max-time 5 --connect-timeout 5 "https://$TARGET/owa/" \
-H "Authorization: Basic dGVzdDp0ZXN0" 2>/dev/null | grep -i "www-authenticate.*basic")
if [[ -n "$BASIC_AUTH" ]]; then
echo " [!] Basic Auth ENABLED — easier brute force vector"
fi
# Check healthcheck endpoint (sometimes exposes version/config)
HEALTH=$(curl -sk --max-time 5 --connect-timeout 5 "https://$TARGET/owa/healthcheck.htm" 2>/dev/null)
if [[ -n "$HEALTH" ]] && echo "$HEALTH" | grep -qi "200 ok"; then
echo " [+] Healthcheck accessible — server status exposed"
fi
Phase 4 — ADFS/Office 365 Recon (hybrid environments)
TARGET_DOMAIN="$1" # e.g., company.com
echo "[*] ADFS/Office 365 recon on $TARGET_DOMAIN"
# Check for ADFS
ADFS_URL="https://sts.$TARGET_DOMAIN/adfs/ls/IdpInitiatedSignOn.aspx"
ADFS_CODE=$(curl -sk -o /dev/null -w "%{http_code}" --max-time 5 --connect-timeout 5 "$ADFS_URL")
[[ "$ADFS_CODE" == "200" || "$ADFS_CODE" == "302" ]] && echo " [+] ADFS: $ADFS_URL (HTTP $ADFS_CODE)"
# Check Office 365 tenant
O365_XML=$(curl -sk --max-time 5 --connect-timeout 5 "https://login.microsoftonline.com/getuserrealm.srf?login=user@$TARGET_DOMAIN&xml=1" 2>/dev/null)
if echo "$O365_XML" | grep -qi "Federated\|Managed"; then
echo " [+] Office 365 tenant: $(echo "$O365_XML" | grep -Eo '<NameSpaceType>\K[^<]+')"
echo " $(echo "$O365_XML" | grep -Eo '<DomainName>\K[^<]+')"
fi
# Autodiscover (leaks internal server names)
AUTODISCOVER=$(curl -sk --max-time 10 --connect-timeout 10 "https://autodiscover.$TARGET_DOMAIN/autodiscover/autodiscover.xml" \
-H "Content-Type: text/xml" \
-d '<?xml version="1.0"?><Autodiscover xmlns="http://schemas.microsoft.com/exchange/autodiscover/outlook/requestschema/2006"><Request><EMailAddress>user@'$TARGET_DOMAIN'</EMailAddress><AcceptableResponseSchema>http://schemas.microsoft.com/exchange/autodiscover/outlook/responseschema/2006a</AcceptableResponseSchema></Request></Autodiscover>' 2>/dev/null)
if echo "$AUTODISCOVER" | grep -qi "server\|internal"; then
echo " [+] Autodiscover response — internal server names leaked"
echo "$AUTODISCOVER" | grep -Eo '(?:<Server>|<InternalRpcClientServer>|<ASUrl>)[^<]+' | head -5
fi
Pitfalls
- NTLM relay requires specific network position. Unless you control a machine the Exchange server can reach, NTLM relay is not exploitable remotely.
- Modern Exchange (Exchange Online, 2019+) blocks Basic Auth by default. Test with Modern Auth (OAuth2) if Basic is blocked.
- Account lockout policies vary. Test with a single known-bad password before spraying.
- ADFS is NOT Exchange. ADFS is a separate service with its own attack surface (SAML, WS-Trust).
Verification
- NTLM Type-2 MUST decode to reveal at minimum DNS Domain Name and NetBIOS Domain Name.
- OWA version MUST be extracted from
X-OWA-Versionheader. - Password spray surface: confirm NO rate limiting (5 rapid attempts all return the same HTTP code).
- Autodiscover MUST return internal server names (not just external URLs).
- Document: Exchange version, AD domain, NetBIOS name, computer names, rate limiting status.
Related Skills
password-spray-methodology— Universal password spray pipeline across all protocols + error code differentials
Related Skills
taste-skill
116.8kAgent skills that give coding agents taste: art direction, reference images and a browser check instead of generic AI pages
Agent-Reach
112.3kGive your AI agent eyes to see the entire internet. Read & search Twitter, Reddit, YouTube, GitHub, Bilibili, XiaoHongShu — one CLI, zero API fees.
headroom
79.8kCompress tool outputs, logs, files, and RAG chunks before they reach the LLM. 20% fewer tokens for coding agents, 60-95% fewer tokens for JSON, same answers. Library, proxy, MCP server.
ai-job-search
72.1kThe job search that runs on your machine. AI job application framework built on Claude Code: evaluate postings, tailor CVs, write cover letters, prep interviews. Fork it and own it.
Languages
Trust signals
From repository metadata: license, adoption, age and documentation. Not a code audit — see the Safety scan above for what the skill file itself contains.
