AGENT SKILL

opsec

Operational security management — traffic shaping, scan rate limiting, source IP management, tool signature avoidance, evidence handling, anti-detection patterns.

License 许可
MIT
最近更新
2026-06-15
Stars
★ 0
Forks
0

01 前言 Front Matter

name
opsec
metadata
{ "tags": "opsec, stealth, rate-limit, user-agent, scope-check, evidence-handling, anti-detection", "subdomain": "opsec", "when_to_use": "OPSEC check, rate limit, stealth, detection avoidance, scan timing, user-agent, scope check, evidence handling, clean up, anti-detection", "mitre_attack": "T1562, T1070, T1036" }
description
Operational security management — traffic shaping, scan rate limiting, source IP management, tool signature avoidance, evidence handling, anti-detection patterns.
allowed-tools
Bash Read

03 SKILL.md 正文

Operational Security (OPSEC) Knowledge Base

OPSEC ensures the red team engagement remains covert, controlled, and within authorized scope. Poor OPSEC burns the engagement — detected scans alert the blue team, taint findings, and waste client resources. This skill applies across all recon phases.

1. Core OPSEC Principles

The OPSEC Mindset

  1. Every packet is a signal — assume the target has IDS/IPS, SIEM, and SOC analysts
  2. Minimize footprint — collect only what you need, no more
  3. Blend with normal traffic — timing, volume, and patterns should look legitimate
  4. Know your tools' signatures — every scanner has a fingerprint
  5. Document everything — if you can't prove it was authorized, it wasn't

Engagement Scope Awareness

Before ANY active operation:

  • [ ] Written Rules of Engagement (ROE) on file
  • [ ] Target IP ranges and domains explicitly listed
  • [ ] Out-of-scope assets clearly identified
  • [ ] Testing window defined (dates, times)
  • [ ] Emergency contact for the client (for accidental impact)
  • [ ] Get-out-of-jail letter accessible

2. Network OPSEC

Scan Rate Limiting

Target Type Recommended Rate Timing Flag
Production web server 5-10 req/sec nmap -T2
Internal network 50-100 req/sec nmap -T3
Development/staging 100+ req/sec nmap -T4
High-security target 1-2 req/sec nmap -T1
WAF-protected 1-5 req/sec Custom delays

Traffic Shaping

# nmap with specific rate limiting
nmap -sS --max-rate 10 --max-retries 1 -p 80,443 <target>

# ffuf with rate limiting
ffuf -u https://<target>/FUZZ -w wordlist.txt -rate 5

# nuclei with rate limiting
nuclei -u https://<target> -rl 5 -c 2

# curl with delay between requests
for url in $(cat urls.txt); do
    curl -s -o /dev/null -w "%{http_code} $url\n" "$url"
    sleep 2
done

Timing Considerations

  • Business hours (9am-5pm target timezone): Higher baseline traffic → scanning blends in
  • Weekends/holidays: Lower baseline → scans stand out more
  • Maintenance windows: If known, ideal for aggressive scanning
  • Burst vs sustained: Short bursts with long pauses are less detectable than sustained scanning

3. HTTP OPSEC

User-Agent Management

Important: User-Agent strings become stale quickly. Always use current browser version strings that match real-world traffic at the time of engagement. Check your own browser's UA or query a live UA database before starting.

# Step 1: Get a current, real User-Agent from your own browser or a live source
# Option A: Copy from your browser's DevTools (Network tab → Request Headers)
# Option B: Use a curated list — update version numbers to match current releases

# Step 2: Build a rotation list with CURRENT versions
# Template — replace <CHROME_VER> and <FIREFOX_VER> with latest stable versions
UA_LIST=(
    "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/<CHROME_VER> Safari/537.36"
    "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/<CHROME_VER> Safari/537.36"
    "Mozilla/5.0 (X11; Linux x86_64; rv:<FIREFOX_VER>) Gecko/20100101 Firefox/<FIREFOX_VER>"
    "Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:<FIREFOX_VER>) Gecko/20100101 Firefox/<FIREFOX_VER>"
)
UA="${UA_LIST[$RANDOM % ${#UA_LIST[@]}]}"

curl -s -A "$UA" https://<target>/

# ffuf with custom user agent
ffuf -u https://<target>/FUZZ -w wordlist.txt -H "User-Agent: $UA"

Header Hygiene

# Avoid tool-specific headers that reveal scanner identity
# BAD: Default tool user agents
# - "Nmap Scripting Engine"
# - "nikto"
# - "gobuster"
# - "sqlmap"

# GOOD: Mimic real browser headers (use current browser versions!)
curl -s https://<target>/ \
    -H "User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 Chrome/<CURRENT_VER> Safari/537.36" \
    -H "Accept: text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,*/*;q=0.8" \
    -H "Accept-Language: en-US,en;q=0.9" \
    -H "Accept-Encoding: gzip, deflate, br"

4. Tool Signature Awareness

Common IDS/WAF Signatures

Tool Detection Signature Mitigation
nmap SYN scan pattern, probe ordering Use -T2, --data-length
nikto Default User-Agent, predictable paths Custom UA, selective tuning
sqlmap Parameter tampering patterns Not applicable to recon phase
ffuf Rapid sequential requests Rate limiting (-rate)
nuclei Template-specific payloads Rate limiting (-rl), selective templates
gobuster Sequential path enumeration Randomize wordlist, rate limit

Reducing Scanner Fingerprint

# nmap — add random data to packets
nmap -sS --data-length 24 -T2 <target>

# nmap — randomize host order (for multi-target)
nmap -sS --randomize-hosts -iL targets.txt

# nmap — spoof source port (use common ports)
nmap -sS -g 53 <target>  # Appear as DNS traffic
nmap -sS -g 80 <target>  # Appear as HTTP traffic

5. Source Management

IP Awareness

  • Know your egress IP: curl -s ifconfig.me
  • Single source: All scans originate from the sandbox — the target will see one IP
  • VPN/proxy considerations: If engagement allows, rotate exit nodes
  • Cloud instances: Ephemeral cloud VMs provide disposable IPs

DNS OPSEC

# Use public resolvers to avoid leaking internal DNS queries
dig @8.8.8.8 <target> A +short
dig @1.1.1.1 <target> A +short

# Don't use target's own DNS servers for recon queries
# (they may log all queries from unknown sources)

6. Evidence & Data Handling

Engagement Documentation

Every action should be logged:

| Timestamp (UTC) | Action | Target | Tool | Justification |
|-----------------|--------|--------|------|---------------|
| <YYYY-MM-DD HH:MM> | SYN scan top 1000 | 10.0.1.50 | nmap | Passive recon identified as primary web server |
| <YYYY-MM-DD HH:MM> | Dir fuzzing | api.example.com | ffuf | Port 443 open, REST API suspected |

Data Classification

  • Workspace files: All scan output goes to the engagement directory — treat as engagement artifacts
  • Credentials found: NEVER store in plaintext — immediately document and encrypt
  • PII discovered: Note existence, do not exfiltrate — document for client
  • Client data: Handle per engagement contract data handling requirements

Clean-Up Protocol

After engagement:

  • [ ] All tools stopped
  • [ ] No persistent connections to target
  • [ ] Scan data secured per client requirements
  • [ ] Temporary files cleaned from sandbox
  • [ ] No data left on third-party services (Shodan searches, etc.)

7. Scope Enforcement

Automated Scope Checking

# Before scanning, verify target is in scope
SCOPE_FILE="scope.txt"
TARGET="10.0.1.50"

if grep -q "$TARGET" "$SCOPE_FILE" 2>/dev/null; then
    echo "IN SCOPE — proceed"
else
    echo "WARNING: $TARGET not found in scope file!"
    echo "Verify before proceeding."
fi

Scope Boundaries

  • IP ranges: Only scan IPs explicitly listed in scope document
  • Domains: Only enumerate subdomains of authorized root domains
  • Cloud resources: Only test resources confirmed as client-owned
  • Third-party services: DO NOT test shared infrastructure (CDNs, SaaS platforms)
  • Physical scope: Only applicable if physical penetration test is authorized

Accidental Out-of-Scope

If you accidentally touch an out-of-scope system:

  1. Stop immediately
  2. Document the incident (timestamp, target, action taken)
  3. Notify the engagement lead
  4. Do NOT attempt to "clean up" evidence — this makes it worse

8. Detection Indicators to Monitor

Signs You've Been Detected

  • Connection resets: Target suddenly dropping connections → firewall rule added
  • Rate limiting: 429 responses where 200s were before → WAF triggered
  • IP block: Timeouts on previously responsive hosts → IP banned
  • Honeypot indicators: Unusually easy targets, too-good-to-be-true services
  • Tarpit responses: Extremely slow responses → deliberately slowing scanner

Response to Detection

  1. Pause all scanning — at least 30 minutes
  2. Assess what triggered detection — review recent scan patterns
  3. Reduce scan rate by 50-75%
  4. Consider different approach — passive only for a period
  5. Document the detection event for engagement report

9. OPSEC Checklist (Pre-Engagement)

  • [ ] ROE document reviewed and understood
  • [ ] Scope file created at scope.txt
  • [ ] Emergency client contact accessible
  • [ ] Egress IP noted: curl -s ifconfig.me
  • [ ] Tool user agents configured (non-default)
  • [ ] Scan rate limits configured per target type
  • [ ] Output directory structure created
  • [ ] Evidence logging template ready
  • [ ] Clean-up protocol documented

04 同仓库更多 Skills