Skip to main content
Press slash or control plus K to focus the search. Use the arrow keys to navigate results and press enter to open a threat.
Reconnecting to live updates…

GitHub’s Fake Engagement Problem Is Hiding in Plain Sight

0
Medium
Published: Wed May 20 2026 (05/20/2026, 21:53:22 UTC)
Source: Reddit ThreatIntel

Description

This threat describes a widespread fake engagement problem on GitHub where bot accounts artificially inflate star counts on repositories, primarily targeting low-quality AI tools and executor software. A detection tool named phantomstars identifies these campaigns by analyzing account creation dates, activity patterns, and coordinated timing of star events. The fake engagement distorts the credibility signals developers rely on to evaluate and recommend software. No direct exploitation or vulnerability is described, but the manipulation impacts trust and decision-making in the open source ecosystem.

Reddit Discussion

00
This Reddit post has been deleted. Content shown was captured before removal.

Turns out: very visible. Yesterday's scan found 185 out of 185 engagers on a single repo were bots. Not 90%. Not "mostly suspicious". Every single one. The repo had zero legitimate stars.

What I built

phantomstars is a Python tool that runs daily via GitHub Actions (free, no servers):

  1. Scrapes GitHub Trending and searches for repos created in the last 7 days with sudden star spikes
  2. Pulls star and fork events from the last 24 hours per repo
  3. Bulk-fetches every engager's profile via the GraphQL API (account creation date, follower counts, repo history)
  4. Scores each account on a weighted model: account age (35%), profile completeness (30%), repo patterns (25%), activity history (10%)
  5. Detects coordinated campaigns using timestamp clustering and union-find: groups of 4+ suspicious accounts that engaged within a 3-hour window
  6. Files an issue directly on the targeted repo so the maintainer knows what's happening

Campaign IDs are deterministic SHA-256 fingerprints of the sorted member set, so the same group of bots gets the same ID across runs. You can track a farm across multiple days even as individual accounts get suspended.

What the pattern actually looks like

It's remarkably consistent. A fake engagement campaign in the raw data:

  • 40-200 accounts, all created within the same 1-2 week window
  • Zero original repositories, or only forks they never touched
  • No bio, no location, no followers, no following
  • All of them starring the same repo within a 90-minute window
  • The target repo usually has a name implying it's a tool, hack, executor, or generator

Today's scan: 53 active campaigns across 3,560 accounts profiled. 798 classified as likely_fake. The repos being targeted are mostly low-quality AI tools and "executor" software that needs manufactured credibility fast.

Notifying the affected repo

When a repo hits a 40%+ fake engagement ratio or a campaign is detected, phantomstars opens an issue on that repo with the full suspect table: account logins, creation dates, composite scores, campaign membership. The maintainer sees it in their own issue tracker without having to find this project first.

Worth noting: a lot of these repos have issues disabled, which is a red flag on its own. Those get skipped silently.

Why I built this

Stars are how developers decide what to evaluate, what to depend on, what to recommend. When that signal is bought, it affects real decisions downstream. This started as curiosity about how measurable the problem was. The answer was more measurable than I expected.

It's part of broader research into AI slop distribution at JS Labs: https://labs.jamessawyer.co.uk/ai-slop-intelligence-dashboards/

The fake engagement problem and the AI content quality problem are really the same problem. Fake stars are the distribution layer that gets garbage in front of real users.

All open source. The data is append-only JSONL committed back to the repo after every run, queryable with jq.

Repo: https://github.com/tg12/phantomstars

Findings are probabilistic, false positives exist, the README explains the full scoring model. If your account shows up and you're a real person, there's a false positive process.

Questions welcome on the detection approach, GraphQL batching, or campaign ID stability.

AI-Powered Analysis

Machine-generated threat intelligence

AILast updated: 05/24/2026, 18:55:16 UTC

Technical Analysis

The threat involves coordinated fake engagement campaigns on GitHub, where large groups of newly created bot accounts with minimal profile information star the same repository within a short time window. The phantomstars tool uses GitHub Actions and the GraphQL API to detect these campaigns by scoring accounts on factors such as account age, profile completeness, repository activity, and timing clusters. When a campaign is detected, an issue is automatically filed on the targeted repository to notify maintainers. This manipulation inflates star counts, misleading users about the quality and popularity of repositories, especially low-quality AI-related projects.

Potential Impact

The impact is primarily reputational and informational, as fake stars can mislead developers into trusting or depending on low-quality or potentially harmful software. This undermines the integrity of GitHub's social signals used for software evaluation. There is no indication of direct exploitation or compromise of systems. The problem affects the open source community's ability to discern genuine project popularity and quality.

Mitigation Recommendations

No official patch or fix is applicable as this is a manipulation of social signals rather than a software vulnerability. The phantomstars tool provides a detection and notification mechanism to alert repository maintainers of suspicious engagement. Repository maintainers can monitor issues filed by such tools and consider disabling stars or taking other repository-level actions if fake engagement is confirmed. Users should remain cautious about relying solely on star counts for evaluating repositories.

Pro Console: star threats, build custom feeds, automate alerts via Slack, email & webhooks.Upgrade to Pro

Technical Details

Source Type
reddit
Subreddit
ThreatIntelligence+threatintel+websecurityresearch
Reddit Score
0
Discussion Level
minimal
Content Source
reddit_link_post
Post Type
link
Domain
null
Newsworthiness Assessment
{"score":32,"reasons":["external_link","established_author"],"isNewsworthy":true,"foundNewsworthy":[],"foundNonNewsworthy":[]}
Has External Source
true
Trusted Domain
false

Threat ID: 6a13498fa5ae1af1aab69491

Added to database: 5/24/2026, 6:55:11 PM

Last enriched: 5/24/2026, 6:55:16 PM

Last updated: 5/24/2026, 10:09:41 PM

Views: 3

Community Reviews

0 reviews

Crowdsource mitigation strategies, share intel context, and vote on the most helpful responses. Sign in to add your voice and help keep defenders ahead.

Sort by
Loading community insights…

Want to contribute mitigation steps or threat intel context? Sign in or create an account to join the community discussion.

Actions

PRO

Updates to AI analysis require Pro Console access. Upgrade inside Console → Billing.

Please log in to the Console to use AI analysis features.

Need more coverage?

Upgrade to Pro Console for AI refresh and higher limits.

For incident response and remediation, OffSeq services can help resolve threats faster.

Latest Threats

Breach by OffSeqOFFSEQFRIENDS — 25% OFF

Check if your credentials are on the dark web

Instant breach scanning across billions of leaked records. Free tier available.

Scan now
OffSeq TrainingCredly Certified

Lead Pen Test Professional

Technical5-day eLearningPECB Accredited
View courses