Integration

Connect GitHub to Polaris

Fine-grained tokens let you hand over three repositories instead of an account, which is the whole reason to use them here.

What connecting buys

Connecting GitHub gives a Polaris worker the repositories, issues and pull requests inside a fine-grained access token's scope. Polaris verifies the token against GitHub before storing it and shows the account login it came back with. Scope the token to the repositories workers need and choose read-only permissions where writing is not part of the job.

Credential
Fine-grained access token
Verified live
GitHub returns the account login
Scope control
Per repository, at GitHub

Scope the token, not the trust

GitHub's fine-grained tokens are the reason this connection is comfortable to grant. You pick the repositories, you pick the permissions, and the token cannot reach anything you left out. A worker that only writes documentation does not need write access to production infrastructure, and a token is where that decision gets enforced.

Polaris checks the token by asking GitHub whose it is, then shows you that login in the connect panel. If the login surprises you, that is the check working.

What the connection gives a worker

Everything below is bounded by the token you created. Narrow the token and this list narrows with it.

  • The repositories you named

    Not the organisation, not every repo the creator can see. A fine-grained token lists its repositories explicitly and that list is the worker's world.

  • Issues and pull requests

    The raw material for triage write-ups, release notes and the summary of what changed this week that nobody has time to assemble.

  • Whatever the permissions allow

    Read-only stays read-only inside Polaris. There is no privilege escalation path in the product, because the credential is used as given.

  • Context to write from

    A technical writer working from real code and real pull-request discussion produces documentation that matches the repository rather than the intention.

Scoping decisions worth making before you paste

Polaris cannot make these for you. GitHub is where they are enforced.

DecisionSafe defaultWhen to widen it
Which repositoriesThe specific repos workers are briefed onNever widen speculatively; add repos when a task needs one
Read or writeRead-onlyOnly if you intend workers to open issues or pull requests
Which accountA machine account or a member with matching accessAvoid keys made on an admin account entirely
ExpiryThe shortest expiry your workflow toleratesLonger only where re-authorizing is genuinely disruptive

Workers and work this connection feeds

Questions people ask

+Does Polaris need admin access to my GitHub organisation?

No. A fine-grained personal access token scoped to specific repositories is what the catalog asks for, and it is the right answer. There is no GitHub App install, no organisation-wide permission request and no reason to hand over more than the repositories in play.

+Can an AI worker push code or merge a pull request?

Only if the token you created allows it. Permissions are set at GitHub, and Polaris uses the credential exactly as issued. A read-only token stays read-only no matter what a worker is asked to do.

+How does Polaris verify a GitHub token?

It calls GitHub's user endpoint with the token before storing anything. A rejected token produces GitHub's status code as an error and nothing is written. A valid one comes back with the account login, which the connect panel shows you.

+Is GitHub a good connection for a non-engineering worker?

Sometimes. A technical writer or a project coordinator reading issues and pull requests is doing legitimate work with it. In those cases give read-only access to the specific repositories they write about, and nothing else.

Your next hire takes 60 seconds.

The software is free — unlimited people, tasks, workstreams and docs. You pay only for work an AI worker actually delivers, itemised by the hour.

Get started free

Last checked .