GitHub Copilot pull-request ads have been removed after developers discovered promotional messages injected into their pull requests, triggering immediate backlash across developer communities. On March 30, 2026, GitHub reversed the feature following complaints that the tips resembled unwanted advertising rather than helpful guidance.
Key Takeaways
- GitHub removed Copilot tips from pull request comments after developers flagged them as ads.
- Over 11,000 pull requests contained the Raycast extension promotion before removal.
- A programming logic bug caused tips meant for Copilot-created PRs to appear in human-created ones.
- The issue affected Copilot Free, Pro, and Pro+ users but not business or enterprise accounts.
- GitHub denied intentional advertising, framing the incident as a technical error, not a monetization strategy.
What Happened With GitHub Copilot Pull-Request Ads
Developers spotted a message appearing in pull request comments: “⚡ Quickly spin up Copilot coding agent tasks from anywhere on your macOS or Windows machine with Raycast.” The phrase, complete with markdown headers reading “START COPILOT CODING AGENT TIPS,” looked and felt like a promotional ad for Raycast, a third-party productivity app. The discovery sparked immediate concern among developers who viewed pull requests as sacred—a space for code review and collaboration, not marketing. Reports indicated that over 11,000 pull requests contained this exact promotional message, with some sources claiming as many as 1.5 million affected PRs across the platform.
The incident highlights growing developer frustration with AI-generated “slop” infiltrating previously clean workflows. Pull requests represent the heart of collaborative development, where code quality and clarity matter most. Injecting promotional content into this context felt like a violation of trust, regardless of whether GitHub intended it as advertising or not.
GitHub’s Official Explanation: Bug, Not Ads
Martin Woodward, VP of Developer Relations at GitHub, addressed the backlash directly. “We identified a programming logic issue with a GitHub Copilot coding agent tip that surfaced in the wrong context within a pull request comment. We have removed agent tips from pull request comments moving forward,” Woodward stated. GitHub also emphasized that “GitHub does not and does not plan to include advertisements in GitHub”.
The root cause was a technical error: tips designed to appear only in pull requests created by Copilot itself were appearing in human-created pull requests when developers invoked Copilot to edit code. The feature had been live since around May 2025, when GitHub released its Copilot coding agent, but the bug went unnoticed until developers spotted the Raycast promotion at scale. The removal came swiftly—within days of public complaints on social media and developer forums like Hacker News.
Why Developers Reacted So Strongly
The backlash reflects deeper anxiety about AI creeping into spaces where developers expect full control. Pull requests are where code gets scrutinized, discussed, and approved before merging into production. Unexpected promotional content in this context felt invasive, even if unintentional. The fact that Copilot tips appeared with emoji and marketing language made them feel less like technical guidance and more like sponsored content.
This incident also sits within a broader trend of AI-generated content appearing in unexpected places—from Windows 11 recommendations to other productivity tools. Developers, already skeptical of AI quality and reliability, saw GitHub Copilot pull-request ads as another example of AI “slop” degrading the user experience rather than improving it. The speed of GitHub’s response—removing the feature entirely rather than refining it—suggests the company understood the severity of developer sentiment.
What This Means for Copilot Users Going Forward
The removal of pull-request tips does not affect Copilot’s core functionality. Free, Pro, and Pro+ users can still invite Copilot to make changes in pull requests by mentioning @copilot, and the AI will continue to generate code suggestions and edits. What is gone is the promotional tip feature that appeared alongside these edits. Enterprise and business customers were never affected by the bug, as their pull requests did not display the tips.
Developers also retain the option to opt out of using their GitHub data to train Microsoft’s models, preserving some control over how their code contributes to Copilot’s training dataset. The incident serves as a reminder that even well-intentioned AI integrations can backfire if they cross implicit boundaries in developer workflows.
Is GitHub Copilot still available in pull requests?
Yes. GitHub Copilot remains fully functional in pull requests. Developers can still invite the AI to suggest changes and edits using the @copilot mention. What was removed was the promotional “tips” feature that appeared as separate messages. The core coding agent functionality continues unchanged.
Why did Raycast appear in the Copilot tips?
Raycast was promoted in the Copilot tips as a way to access the Copilot coding agent from macOS or Windows machines directly, without opening GitHub. However, the tips appeared in pull requests where they should not have due to a programming logic error, making the promotion look unintended and ad-like.
Can developers block Copilot-generated pull requests?
The incident sparked discussions on platforms like Hacker News and GitHub’s community forums about whether developers should be able to block pull requests created entirely by Copilot or filter out Copilot-generated content. GitHub has not announced any blocking feature, but the conversation reflects ongoing concerns about maintaining developer agency in code review workflows.
GitHub Copilot pull-request ads were a misstep, but the swift removal signals that GitHub is listening to developer feedback. The incident underscores a critical lesson: AI tools must respect the boundaries of existing workflows, especially in spaces like pull requests where trust and clarity are non-negotiable. As AI becomes more embedded in development tools, this tension between automation and control will only grow sharper.
Edited by the All Things Geek team.
Source: TechRadar
