osint-social

Social media reconnaissance skill — platform enumeration, Google dork patterns, cross-platform identity resolution, public data collection without scraping, and digital footprint analysis.

SKILL.md

GitHub

OSINT Social Media Reconnaissance

Social media intelligence from public sources — platform enumeration, Google dorking, cross-platform identity resolution, and digital footprint analysis using only publicly available information and respecting platform terms of service.

Prerequisites

Required Tools

  • Web browser (Firefox DevTools MCP)
  • Terminal with curl/jq
  • No MCP servers required for manual public-data techniques

Recommended (but optional) Tools

Public Social Media Sources

Major Platforms

Platform Public Info Available Typical Use Case
LinkedIn Profile, job history, education, skills, connections (partial) Professional recon
Facebook Name, profile pic, public posts, pages, groups, photos Personal recon
Twitter/X Bio, posts, follows, followers, media, lists Real-time activity monitoring
Instagram Bio, photos, followers (public accounts) Visual recon, location patterns
TikTok Bio, public videos Youth demographic recon
Reddit Post/comment history, subreddit participation Interests, opinions
GitHub Repos, contributions, profile, organizations Technical skills
YouTube Channel, videos, comments, playlists Content interests
Pinterest Boards, pins, interests Lifestyle, design, planning
Medium Articles, responses, publications Professional writing
Substack Newsletters, subscriber counts Thought leadership
Strava Exercise routes, heat maps Location inference
Meetup Groups, events, RSVPs Professional/personal network
Goodreads Books read, reviews Reading preferences
Quora Answers, questions, topics Knowledge areas
Wikipedia Edit history, talk page participation Specialized knowledge
Stack Overflow Answers, reputation, badges Technical expertise
Foursquare Venue tips, check-ins (historical) Location patterns
Eventbrite Events hosted/attended Professional networking

Google Dork Patterns

Basic Google Operators

"exact phrase"        → Exact match search
site:example.com      → Limit to specific site
-inurl:spam           → Exclude URLs containing "spam"
intitle:"profile"     → Pages with "profile" in title
inurl:about           → URLs containing "about"
filetype:pdf          → Specific file types
link:example.com      → Pages linking to example.com
related:example.com   → Similar sites
cache:example.com     → Google cached version (historical)

Social Media Specific Dorks

# LinkedIn
site:linkedin.com/in "Jane Smith" "Portland"
site:linkedin.com/in "Software Engineer" "Jane"
site:linkedin.com/company "Acme Corp" employees
site:linkedin.com/company "Acme Corp" about
site:linkedin.com/pub "Jane Smith" "Portland"

# Facebook
site:facebook.com "Jane Smith" "Portland"
site:facebook.com intitle:"Jane Smith"
site:facebook.com/people "Jane" "Portland" "OR"
site:facebook.com/public/"Jane-Smith"

# Twitter / X
site:twitter.com "Jane Smith" Portland
site:x.com "Jane Smith" Oregon
site:twitter.com intitle:"Jane Smith"

# Instagram
site:instagram.com "jane.smith"
site:instagram.com/p/ "Jane Smith"           # Photo captions

# Reddit
site:reddit.com "Jane Smith"
site:reddit.com/user/"Jane Smith"
site:reddit.com "Smith" subreddit:Portland

# GitHub
site:github.com "Jane Smith"
site:github.com/ "jane-smith" inurl:README
site:github.com/ "email: jane@example.com"   # Look for exposed emails

# YouTube
site:youtube.com "Jane Smith" inurl:channel
site:youtube.com "Jane Smith" inurl:about
site:youtube.com/watch "Jane Smith" intitle:"comment"  # Comments

# Medium
site:medium.com "Jane Smith"
site:medium.com/@jane.smith

# General presence
site:about.me "Jane Smith" Portland
site:linktr.ee "Jane Smith"
site:behance.net "Jane Smith"
site:dribbble.com "Jane Smith"
site:pinterest.com "Jane Smith"
site:goodreads.com "Jane Smith"
site:gravatar.com "Jane Smith"
site:keybase.io "Jane Smith"
site:angel.co "Jane Smith"                   # AngelList
site:crunchbase.com "Jane Smith"             # Startup profiles
site:zoominfo.com "Jane Smith"               # Professional profiles
site:spokeo.com "Jane Smith"                 # Public records aggregator

Advanced Dork Patterns

# Find email addresses
"Jane Smith" "@" "gmail.com" OR "@" "yahoo.com" OR "@" "outlook.com"
"Jane Smith" email
site:linkedin.com/in "Jane" email

# Find resumes (often contain detailed personal info)
intitle:resume "Jane Smith" Portland filetype:pdf
intitle:cv "Jane Smith" filetype:docx
"Jane Smith" resume "Software Engineer" Portland

# Find documents with personal info
"Jane Smith" phone filetype:pdf
"Jane Smith" address filetype:pdf
"Jane Smith" "SSN" - DO NOT USE — illegal search pattern
"Jane Smith" "DOB" filetype:xlsx

# Breach/corpus mentions
"Jane Smith" "haveibeenpwned"
"Jane Smith" "password" filetype:txt           # Potential breach remnants

# Personal websites
"Jane Smith" "my website" OR "personal website"
"Jane Smith" "wix.com" OR "squarespace.com" OR "wordpress.com"

# News mentions
"Jane Smith" "ABC Corp" site:nytimes.com
"Jane Smith" "ABC Corp" site:wsj.com
"Jane Smith" "ABC Corp" site:bloomberg.com
"Jane Smith" "award" OR "promoted" OR "hired"

Step-by-Step Workflows

1. Identity Resolution — Name → Profiles

# Discovery workflow
name = "Jane Smith"
location = "Portland, OR"
employer = "Acme Corp"  # If known

# Phase 1: Google reconnaissance
google_queries = [
    f'"{name}" "{location}"',
    f'"{name}" linkedin',
    f'"{name}" "{employer}"',
    f'"{name}" github',
    f'"{name}" twitter',
    f'"{name}" facebook',
]

# Phase 2: Platform-specific searches
platforms = {
    "linkedin": f"site:linkedin.com/in \"{name}\" \"{location}\"",
    "facebook": f"site:facebook.com \"{name}\" \"{location}\"",
    "twitter": f"site:twitter.com \"{name}\" \"{location}\"",
    "github": f"site:github.com \"{name}\" \"{employer}\"",
}

2. Username → Cross-Platform Mapping

# ── SHERLOCK ── Username search across 400+ social networks ──
# Install: pip install sherlock-project
# GitHub: https://github.com/sherlock-project/sherlock

# Basic usage
sherlock jane.smith123

# Check multiple usernames at once
sherlock jane.smith123 jsmith jane_smith

# Save results to JSON
sherlock jane.smith123 --output /tmp/sherlock_results.json

# Save to CSV for spreadsheet analysis
sherlock jane.smith123 --csv

# Save to Excel
sherlock jane.smith123 --xlsx

# Print only found accounts (clean output)
sherlock jane.smith123 --print-found

# Increase timeout for slow sites (default 60s)
sherlock jane.smith123 --timeout 30

# Use Tor proxy (requires tor to be running)
sherlock jane.smith123 --tor

# Check a single specific site by name
sherlock jane.smith123 --site instagram

# Bypass NSFW site filtering (checks adult platforms too)
sherlock jane.smith123 --nsfw

# Username variation engine — replaces {?} with _ - .
sherlock jane{?}smith     # Checks: jane_smith, jane-smith, jane.smith

# Check from a list of usernames in a file
sherlock --folderoutput /tmp/sherlock/ $(cat usernames.txt)

# Real-world example: find all accounts for a target
sherlock yourusername --print-found --output /tmp/paulie_accounts.json

# Method 2: WhatIsMyName (web)
# navigate_page to: https://whatsmyname.app/
# Enter username → returns matches across 300+ platforms

# Method 3: Manual search
for platform in "github.com" "twitter.com" "instagram.com" "reddit.com" "medium.com" "keybase.io"; do
  echo "Searching $platform/jane.smith123..."
  curl -sI "https://www.$platform/jane.smith123" | head -1
done

Username Variations to Try:

jane.smith
janesmith
jane_smith
jane-smith
j.smith
janesmith123
jane.smith.portland
jane.smith.90210
jane.smith.acme
jsmith
jane_s
itsjanesmith
iamjanesmith
jane_the_smith
ms.jane.smith

3. Email → Cross-Platform Registration Check (Holehe)

Holehe checks if an email address is registered on 100+ platforms without sending a password reset email or alerting the target. It works by exploiting subtle differences in how registration/login pages respond to existing vs. non-existent accounts — the same technique used by federal investigators.

Install: pip install holehe GitHub: https://github.com/megadose/holehe

# Basic usage — check email across all platforms
holehe target@email.com

# Show only platforms where the email IS registered (clean output)
holehe target@email.com --only-used

# Disable password recovery attempts (faster, less detectable)
holehe target@email.com -NP

# CSV output for record-keeping
holehe target@email.com -C

# Increase timeout for slow sites (default: 10s)
holehe target@email.com -T 15

# Check multiple emails at once
holehe target1@email.com target2@email.com --only-used

# Real-world example: confirm a target's email usage
holehe the operator@gmail.com --only-used -NP

What Holehe reveals:

  • Which social platforms the email is tied to (Facebook, Twitter, Instagram, LinkedIn)
  • Which professional/creative platforms (GitHub, Medium, Adobe, Spotify)
  • Which utilities/services (Amazon, PayPal, Discord, Telegram)
  • Dating apps, forums, and industry-specific platforms

Detection evasion:

  • -NP flag skips actual password recovery flows — reduces footprint
  • No email is ever sent to the target
  • Results come from analyzing HTTP response differences, not from triggered notifications
  • The tool is silent — the target has no way to know you checked

Cross-reference with Sherlock:

Email (Holehe) → platforms registered → reveals associated usernames
Username (Sherlock) → cross-platform accounts → confirms identity

4. LinkedIn Profile Reconnaissance

# LinkedIn public profile analysis
# Requires Firefox DevTools MCP

# Step 1: Search Google for the profile
# Step 2: Open LinkedIn profile (may need to be logged in)
# Step 3: Extract public info

linkedin_profile_data = {
    "name": "Jane Smith",
    "headline": "Software Engineer at Acme Corp",
    "location": "Portland, Oregon Area",
    "about": "Full-stack engineer...",  # "About" section
    "experience": [
        {"company": "Acme Corp", "title": "Senior Engineer", "dates": "2020-Present"},
        {"company": "TechCo", "title": "Developer", "dates": "2015-2020"},
    ],
    "education": [
        {"school": "Oregon State University", "degree": "BS CS", "years": "2010-2014"}
    ],
    "skills": ["Python", "JavaScript", "AWS", "Kubernetes"],
    "recommendations": 5,          # Number (counts of recommendations)
    "connections_count": 500,      # May not be visible
    "recent_activity": ["Posted about X...", "Commented on Y..."],  # If public
}

What LinkedIn Reveals (Public without login):

- Full name (may show "Jane S." if privacy settings)
- Headline (current role)
- Location (city/metro area)
- Industry
- Number of connections (500+, 3rd+)
- Recent profile activity
- Profile photo (may be public)
- Featured posts (if public)

**What LinkedIn Reveals (Logged In, Not Connected):**
- More detailed experience timeline
- Education details
- Skills (partial)
- Recommendations count
- About section
- Recent activity feed

**What LinkedIn Reveals (1st Connection):**
- Full profile (except private sections)
- Mutual connections
- Contact info (if shared)
- All skills and endorsements

5. Twitter/X Profile Reconnaissance

# Public Twitter data collection (no auth needed for basic)
# Step 1: Navigate to profile
# Step 2: Extract bio, header, pinned tweet
# Step 3: Review recent public tweets

# Advanced: Search tweets by user
curl -s "https://api.twitter.com/2/tweets/search/recent?\
query=from:janesmith\
&max_results=10" | jq .

# Tweet analysis patterns:
# - Location data from tweet metadata
# - Retweet patterns (affiliations)
# - Reply threads (connections)
# - Liked tweets (interests, may not be public)
# - Lists (categorization by user or others)
# - Follows/Followers (network mapping)

6. Facebook Reconnaissance

Facebook public information (varies by privacy settings):
- Name and profile picture (may be public)
- Cover photo (often public)
- Public posts (varies by user)
- About section (selected fields)
- Friends list (user-dependent)
- Liked Pages (may reveal interests)
- Groups (public group memberships)
- Reviews written (public)
- Events attended (public events)

Facebook Search Techniques:

# Facebook graph search (limited as of 2024)
https://www.facebook.com/search/people/?q=jane%20smith%20portland
https://www.facebook.com/search/posts/?q=jane%20smith
https://www.facebook.com/public/jane-smith

# Facebook's directory page
https://www.facebook.com/directory/people/

7. GitHub Reconnaissance

GitHub public information:
- Profile: Name, bio, location, website, company, orgs
- Repositories: Code, README, issues, PRs
- Contributions: Calendar graph, activity history
- Gists: Code snippets
- Comments: On issues, PRs, commits
- Organizations: Affiliations
- Sponsors: If enabled
- GPG keys: Email addresses

Search patterns:
# Find email in commits
git log --format="%an <%ae>" | sort -u

# Find personal info in repos
# Search code for: API keys, configs, passwords
site:github.com "jane.smith" password
site:github.com "jane.smith" API_KEY
site:github.com "jane.smith" aws_secret_access_key

8. Digital Footprint Compilation

## Digital Footprint: Jane Smith

### Professional Platforms
- [x] LinkedIn: linkedin.com/in/janesmith (500+ connections, active)
- [x] GitHub: github.com/janesmith (150 contributions last year)
- [ ] Medium: Not found
- [x] Stack Overflow: stackoverflow.com/users/12345 (2k reputation, 45 answers)

### Social Platforms
- [x] Twitter/X: @janesmith (2,500 tweets, 1,200 followers)
- [x] Facebook: facebook.com/jane.smith (limited privacy settings)
- [x] Instagram: @jane.smith.pdx (private account — bio only)
- [ ] TikTok: Not found
- [ ] Reddit: Not found with this name/username

### Content Platforms
- [x] YouTube: youtube.com/@janesmith (12 videos, 500 subs)
- [ ] Pinterest: Not found
- [ ] Substack: Not found
- [x] Goodreads: goodreads.com/janesmith (150 books)

### Other
- [x] Keybase: keybase.io/janesmith (verified social proofs)
- [x] About.me: about.me/janesmith
- [x] AngelList: angel.co/u/janesmith

### Breach / Security
- [x] HaveIBeenPwned: Email jane.smith@email.com found in 3 breaches
- [x] DeHashed: 2 records found (name + email)
- [ ] Dark web: Not searched (see osint-threat for methodology)

### Footprint Score: HIGH (13+ platforms, active on multiple)

Common Pitfalls

Privacy Settings

  • PITFALL: Privacy settings change constantly; what was public yesterday may not be today.
  • SOLUTION: Collect and cache data promptly. Note date of collection.
  • WORKAROUND: Check Google cached versions for historical content.

Fake / Impersonation Accounts

  • PITFALL: Multiple accounts for the same name — at least some are impersonators or spam.
  • SOLUTION: Verify identity cross-referencing photo, bio details, follower overlaps.
  • WORKAROUND: LinkedIn verified profiles, Twitter blue check, GitHub verified email.

Platform API Changes

  • PITFALL: Platforms frequently change API access (Twitter/X drastically reduced free API).
  • SOLUTION: Prefer web-based manual methods over API-dependent automation.
  • WORKAROUND: Firefox DevTools MCP for extracting visible web data.

Regional Platform Preferences

  • PITFALL: Subject may use platforms dominant outside the US (VK, WeChat, Telegram).
  • SOLUTION: Consider subject's geographic/cultural context.
  • WORKAROUND: In Russia → search VK; China → WeChat/Weibo; Japan → LINE.

Deleted / Abandoned Accounts

  • PITFALL: Old profiles may still exist but be abandoned, giving a misleading impression.
  • SOLUTION: Check last activity date on each platform.
  • WORKAROUND: Archived content may persist (Internet Archive, Google Cache).

Legal & Ethical Notes

⚠️ WARNING: Social media OSINT is heavily regulated:

  • No Automated Scraping: LinkedIn, Facebook, Instagram, Twitter/X prohibit automated data collection in their ToS
  • No Impersonation: Creating fake profiles to gather intel is fraud and illegal in many states
  • No Deception: Do not connect with targets under false pretenses
  • No Private Data: Do not attempt to access private/restricted content
  • Stalking Laws: Monitoring someone's social media activity for harassment is illegal
  • DPPA: Location data from Strava/FourSquare may reveal home addresses
  • Photo Rights: Do not use profile photos without permission
  • GDPR/ePrivacy: EU subjects have enhanced protections
  • UK Investigatory Powers Act: Broader surveillance restrictions

Permissible Methodology

  • Use Google and search engines to find public profiles
  • View only what is accessible without logging in (or with your own logged-in account viewing public content)
  • Use Google cached pages for historical content
  • Use Internet Archive (Wayback Machine) for historical website content
  • Document everything with timestamps and URLs

What NOT to Do

  • ❌ Do not create fake accounts to friend/follow targets
  • ❌ Do not use automated tools that scrape platforms
  • ❌ Do not access private/restricted content
  • ❌ Do not harvest contact information for spam
  • ❌ Do not share screenshots of private interactions
  • ❌ Do not use information for harassment, stalking, or doxxing

Cross-References

  • security/osint-recon — Full OSINT investigation pipeline
  • security/osint-person — Detailed person investigation with enrichment
  • security/osint-facial — Facial recognition matching across social platform photos
  • security/osint-threat — Breach data analysis for exposed credentials
  • security/osint-redteam — Social engineering vector identification from social data
  • software-development/web-scraping-scrapling — If you need legitimate scraping with proper headers/respect
  • software-development/systematic-debugging — Systematic approach to identity resolution problems

Verification Checklist

  • Google search performed with name + location + employer
  • LinkedIn profile found and analyzed (public view)
  • Google dork patterns executed for each major platform
  • Username search across 10+ platforms
  • GitHub profile checked for email/info exposure
  • Twitter/X profile reviewed for connections and activity
  • Facebook public profile reviewed
  • Digital footprint score assigned
  • Breach databases checked (HIBP, DeHashed)
  • Privacy settings respected — no private content accessed
  • All findings dated with source URLs
  • Legal constraints documented
  • Verification note: Social media data alone is insufficient for identity confirmation