Skip to main content

How to create a Signal agent for GitHub activity

Track stars, forks, issues, pull requests, and comments on public GitHub repositories and turn the developers behind them into lemlist contacts.

With a GitHub Activity signal agent, lemlist watches the public GitHub repositories you choose and creates a contact when someone engages with them: a star, a fork, a new issue, a new pull request, or a comment on an issue or pull request. Use it to reach developers who are exploring, evaluating, or actively using a project in your space, such as your own repository or a competitor's.

Each signal includes the GitHub user's profile, enriched with professional contact data (email, company, role) when available.


Limitations

  • You need a GitHub token connected to your team. Without it, repositories cannot be added.

  • You can track up to 10 repositories per signal agent.

  • You can only track public GitHub repositories.

  • Enrichment depends on the GitHub user having identifiable profile information (email, company, real name). Users who cannot be identified do not become contacts, so many events produce no contact.

  • High-traffic repositories (for example, 10K+ stars) can generate a large volume of signals. Select only the events you need and set a daily limit that fits your budget.

  • Each signal agent is scanned once per hour. If a repository receives a lot of activity between two scans, some events can be missed.

  • Removing a star is not tracked. A contact created from a star stays in lemlist if the person later removes the star.

  • The signal agent identifies signals only up to the daily limit you set. Events beyond that limit on the same day do not create contacts.

  • If the signal agent is paused for insufficient credits, it does not resume automatically when your Signals add-on credits are refilled.


Prerequisites

  • A GitHub token is connected in Settings > Integrations > GitHub.

  • You have the URLs of the public repositories you want to monitor.


Credits

GitHub Activity costs 100 credits per signal identified. If your team has the Signals add-on, credits are taken from the Signals add-on first, then from your generic credits.

The daily limit you set in Identification limit caps the cost. For example, a limit of 5 signals per day costs at most 500 credits daily, 15,000 credits per month.


Phase 1: Select the GitHub Activity signal

  1. Open Signal agents from the left sidebar, then click Create signal agent.

  2. Open the Buyer intent category, or type GitHub in the Search signals, keyword, ... bar.

  3. Select GitHub Activity, then click Next.


Phase 2: Configure your signal

  1. Enter an Agent name.

  2. In Repositories to monitor, paste the URL of each public repository in Repository URLs (one per line), for example https://github.com/vercel/next.js.

  3. Click Add repositories.

    ✅ Verify: each repository appears in the list below with Valid.

  1. In GitHub events to monitor, keep the events you want to track and uncheck the others: Star, Fork, Issue, Pull Request, Comment. All five are selected by default. Issue and Pull Request track newly opened ones, and Comment tracks comments on issues and pull requests.

  1. In Identification limit, set the Maximum number of signals identified per day. Use the slider, pick 5, 20, or 50, or type a value in Enter amount. The maximum cost is shown under This will cost at most:.

  2. Click Next.


Phase 3: Scope

In the Scope step, All contacts is the only option for GitHub Activity. Click Next.


Phase 4: Decide how signals are processed

  1. In Signals processing, choose one option:

    • Review signals manually: review and handle each signal yourself.

    • Auto-create tasks: lemlist creates a task for each signal.

    • Auto-push to campaign: lemlist enrolls the new contacts into a campaign.

  2. If you choose Auto-push to campaign, pick one option in Campaign configuration:

    • AI-powered outreach: lemlist generates a campaign with AI-written copy. Choose a sequence type: Email sequence, LinkedIn first, or Multichannel + Cold call, then follow the steps.

    • Link to an existing campaign: connect the signals to a campaign you already created.

  3. Click Next.


Phase 5: Review and deploy

  1. In Summary, check Billing, Signals processing, and Scope.

  2. Click Deploy agent.

    ✅ Verify: the GitHub Activity signal agent appears in Signal agents.


What happens after you deploy

lemlist scans the tracked repositories once per hour. When one of the selected events comes from a GitHub user who can be identified, lemlist creates a contact and processes it the way you chose in Signals processing. Users who cannot be identified are skipped.


Practical example

A company selling a developer tool tracks its own repository and two competitor repositories. It keeps Issue, Pull Request, and Comment, which show active usage, and unchecks Star on the busiest competitor repository to keep the volume manageable. With Auto-push to campaign selected, developers who can be identified are added as contacts and enrolled in the campaign, so outreach starts while the interest is recent.


Troubleshooting & pitfalls

Issue: No GitHub token is connected on this team. Connect one in Settings > Integrations > GitHub

  • Root cause: your team has no GitHub token, so lemlist discards the repositories you add (for example 1 repository discarded). The configuration step also shows Connect your GitHub token to monitor repositories.

  • Fix: click Go to Integrations, or open Settings > Integrations > GitHub, connect a token, then add the repositories again.

Issue: A repository had new activity, but no contacts were created

  • Root cause: the GitHub users cannot be identified, the event type is unchecked in GitHub events to monitor, the daily identification limit is reached, or the next hourly scan has not run yet.

  • Fix: wait at least one hour after the activity. Check that the event type is selected, check the daily count in Identification limit (for example Today: 5 / 5 identified), and raise the limit if needed.

Issue: Too many signals from one repository

  • Root cause: high-traffic repositories generate a lot of stars and comments.

  • Fix: uncheck Star or Comment for that agent, or lower the daily identification limit.

Issue: The signal agent stopped after I added credits

  • Root cause: a signal agent paused for insufficient credits does not resume automatically when the Signals add-on credits are refilled.

  • Fix: open the signal agent in Signal agents and check its status after adding credits.

For other cases where a signal agent shows no results, see Fix: Signal agents not returning results.


If your case is different

Did this answer your question?