Skip to content
Irregularpedia Logo Irregularpedia

Automated Features

The bot runs several background processes that automatically detect, process, and respond to content shared in Signal chats. These features work without explicit commands - they trigger automatically when relevant content is detected.

Per-Room Controls

Section titled “Per-Room Controls”

All automated features can be enabled/disabled per room using !room commands:

!room                      # View all current settings
!room <feature> on/off     # Toggle a specific feature
FeatureCommandDefaultDescription
Auto-Scan!room autoscan on/offONVirus scan & archive attachments
News Summarizer!room news on/offONAuto-post news articles to forum
Tracker Removal!room tracker on/offONStrip tracking params from URLs
Meme Reactions!room memereact on/offONAuto-respond to trigger words
Bot Master Switch!room bot on/offONDisable ALL bot features

Auto-Scan & File Archiving

Section titled “Auto-Scan & File Archiving”

When files are shared in a chat, the bot automatically:

  1. Virus Scans using ClamAV antivirus
  2. Archives clean files to MinIO S3 storage
  3. Reports results with file links

Supported File Types

Section titled “Supported File Types”

Scanned & Archived:

  • Documents: PDF, DOC, DOCX, XLS, XLSX, PPT, PPTX
  • Text: TXT, MD, CSV, JSON, XML
  • Archives: ZIP, RAR, 7Z, TAR, GZ

Skipped (not scanned):

  • Images: JPG, PNG, GIF, WEBP (handled by Signal)
  • Videos: MP4, MOV, AVI (too large)
  • Audio: MP3, WAV, OGG

Example Output

Section titled “Example Output”
🛡️ Auto-Scan Results


📄 report.pdf
✅ Clean - No threats detected
📦 Size: 2.5 MB
📁 Archived: s3.irregular.chat/files/...


📄 data.xlsx
✅ Clean - No threats detected
📁 Archived: s3.irregular.chat/files/...

Threat Detection

Section titled “Threat Detection”

If a virus is detected:

  • File is NOT archived
  • Warning sent privately to uploader via DM
  • No public alert (to avoid embarrassment)

Toggle Per-Room

Section titled “Toggle Per-Room”
!room autoscan off   # Disable auto-scanning for this room
!room autoscan on    # Re-enable auto-scanning

News Article Summarizer

Section titled “News Article Summarizer”

When news URLs are shared, the bot automatically:

  1. Detects news domains (400+ news sites supported)
  2. Scrapes article content using multi-strategy extraction
  3. Generates AI summary using OpenAI
  4. Posts to forum (forum.irregularchat.com)

How It Works

Section titled “How It Works”
User shares: https://arstechnica.com/article/...


Bot automatically:
1. Extracts article title, content
2. Generates 2-3 sentence AI summary
3. Posts to forum with link back to Signal discussion
4. Replies in chat with forum link

Supported News Sources

Section titled “Supported News Sources”

The bot recognizes 400+ news domains including:

Major News: NYTimes, Washington Post, BBC, CNN, Reuters, AP News, NPR, Guardian

Tech News: Ars Technica, TechCrunch, The Verge, Wired, Engadget, ZDNet

Business: Bloomberg, Forbes, WSJ, CNBC, Business Insider

Defense/Military: Defense One, Breaking Defense, C4ISRNet, War on the Rocks, Defense News

Science: Nature, Science, Scientific American, MIT Technology Review

Excluded Domains

Section titled “Excluded Domains”

The bot skips (doesn’t auto-post):

  • Social media (Instagram, TikTok, YouTube, Twitter)
  • Community sites (irregularpedia.org, irregularchat.com)
  • File hosting (Dropbox, Google Drive, GitHub)

Toggle Per-Room

Section titled “Toggle Per-Room”
!room news off    # Disable news auto-posting for this room
!room news on     # Re-enable news auto-posting

URL Tracker Removal

Section titled “URL Tracker Removal”

When URLs are shared, the bot automatically strips tracking parameters and alerts on suspicious domains.

Tracking Parameters Removed

Section titled “Tracking Parameters Removed”

Google Analytics:

  • utm_source, utm_medium, utm_campaign, utm_term, utm_content
  • gclid, gclsrc, dclid, gbraid, wbraid

Facebook/Meta:

  • fbclid, fb_action_ids, fb_ref, fb_source
  • mibextid, sfnsn, _rdr

Instagram:

  • igsh, igshid, ig_rid

Other Platforms:

  • LinkedIn: trk, trkInfo, lipi, rcm
  • Twitter: twclid, s, t
  • TikTok: tt_medium, tt_content, _r
  • YouTube: si, pp, feature

Example

Section titled “Example”
User shares: https://instagram.com/p/ABC123/?igsh=xyz&utm_source=copy


Bot responds:
🔗 Clean URL: https://instagram.com/p/ABC123/
📊 Removed: igsh, utm_source

Suspicious Domain Alerts

Section titled “Suspicious Domain Alerts”

The bot alerts on URLs from high-risk TLDs:

TLDCountryRisk
.ru, .suRussiaHigh
.cn, .com.cnChinaHigh
.irIranHigh
.kpNorth KoreaHigh
.tk, .ml, .gaFree TLDsPhishing

Example Alert:

⚠️ Security Alert


URL from suspicious domain detected:
🔗 example.ru


Country: Russia
Risk: High-risk TLD often associated with malicious content


Proceed with caution. Do not enter credentials.

Toggle Per-Room

Section titled “Toggle Per-Room”
!room tracker off    # Disable tracker removal for this room
!room tracker on     # Re-enable tracker removal
Section titled “Social Media Link Processing”

When social media URLs are shared, the bot:

  1. Cleans tracking params (igsh, fbclid, etc.)
  2. Identifies content type (Reel, Tweet, Video, etc.)
  3. Saves to database for later discovery via !links

Supported Platforms

Section titled “Supported Platforms”
  • Instagram (Posts, Reels, Stories)
  • TikTok (Videos)
  • Twitter/X (Tweets, Threads)
  • YouTube (Videos, Shorts, Live)
  • Facebook (Posts, Videos, Reels)
  • Reddit (Posts, Comments)
  • LinkedIn (Posts, Articles)

Discovery Command

Section titled “Discovery Command”
!links -social       # View recent social media links
!links -social -t 7d # Last 7 days
!links -c -social    # This group only

Git Repository Detection

Section titled “Git Repository Detection”

When GitHub/GitLab URLs are shared, the bot:

  1. Detects repository URLs
  2. Extracts repo metadata (name, description, stars)
  3. Generates AI summary of the project
  4. Saves to database for discovery

Example

Section titled “Example”
User shares: https://github.com/anthropics/claude-code


Bot responds:
📦 Repository: anthropics/claude-code
⭐ Stars: 12.5k
📝 AI Summary: CLI tool for interacting with Claude AI...


Saved to !links -git

Discovery Command

Section titled “Discovery Command”
!links -git          # View shared git repositories
!links -git -t 30d   # Last 30 days

Meme Auto-Reactions

Section titled “Meme Auto-Reactions”

The bot can automatically respond with memes when trigger words are detected.

How It Works

Section titled “How It Works”
  1. Admin sets trigger words for a meme: !meme trigger 42 potato spud
  2. When someone says “potato” or “spud” in chat
  3. Bot has a chance to respond with meme #42

Probability & Cooldown

Section titled “Probability & Cooldown”
  • Not every trigger fires (probability-based)
  • Cooldown prevents spam (can’t trigger same meme twice in short period)
  • Only works in group chats

Toggle Per-Room

Section titled “Toggle Per-Room”
!room memereact off   # Disable auto meme reactions
!room memereact on    # Re-enable auto meme reactions

Identity Fingerprint Polling (Safety Number Detection)

Section titled “Identity Fingerprint Polling (Safety Number Detection)”

The bot polls signal-cli’s identity store every 5 minutes to detect safety number changes that are silently auto-trusted.

How It Works

Section titled “How It Works”
  1. Calls listIdentities via JSON-RPC (returns ~2300+ identities with fingerprints)
  2. Compares each fingerprint against the stored value in known_identity_fingerprints table
  3. If a fingerprint changed AND the user is in a community group:
    • Triggers the safety number verification flow
    • Adds user to Entry/INDOC with 24h deadline
    • Notifies admins in Actions chat
    • Logs the change in identity_change_log

Why This Exists

Section titled “Why This Exists”

Signal-cli runs with --trust-new-identities always, which silently auto-trusts new identity keys without throwing UntrustedIdentityException. The exception-based detection missed many safety number changes. Polling closes this gap.

Monitoring

Section titled “Monitoring”
Terminal window
# Check polling logs
docker logs signal-bot-selfhosted-signal-bot-1 2>&1 | grep IDENTITY_POLL


# Check detected changes
docker exec signal-bot-postgres psql -U signal_bot -d signal_bot \
  -c "SELECT * FROM identity_change_log ORDER BY detected_at DESC LIMIT 5;"

Configuration

Section titled “Configuration”
  • Interval: 5 minutes (hardcoded as IDENTITY_POLL_INTERVAL)
  • First poll: 30 seconds after startup (seeds database without alerting)
  • Scope: Only triggers for users who are members of at least one community group
  • Cannot be disabled per-room (runs globally)

Feature Summary Table

Section titled “Feature Summary Table”
FeatureTriggerOutputToggle
Auto-ScanFile attachmentVirus report + archive link!room autoscan
News SummarizerNews URL sharedForum post + AI summary!room news
Tracker RemovalAny URL with trackingClean URL reply!room tracker
Suspicious DomainHigh-risk TLDSecurity alertAlways on
Social MediaSocial URL sharedClean URL + saved!room tracker
Git RepoGitHub/GitLab URLAI summary + savedAlways on
Meme ReactionsTrigger wordsMeme GIF!room memereact
Identity PollingEvery 5 minSafety number flowAlways on

Configuration

Section titled “Configuration”

These environment variables control automated features:

Terminal window
# Forum posting
DISCOURSE_URL=https://forum.irregularchat.com
DISCOURSE_API_KEY=xxx
DISCOURSE_API_USERNAME=bot
DISCOURSE_CATEGORY_ID=5


# File archiving
MINIO_ENDPOINT=http://minio:9000
MINIO_ACCESS_KEY=xxx
MINIO_SECRET_KEY=xxx
MINIO_BUCKET=irregularchat
MINIO_PUBLIC_URL=https://s3.irregular.chat


# AI summaries
OPENAI_API_KEY=sk-xxx


# Community domains to skip (comma-separated)
COMMUNITY_DOMAINS=irregularpedia.org,irregularchat.com,irregular.chat
Section titled “Related Pages”