GitHub vs Linear for AI agents
As of 16 Oct 2026, GitHub gets an AI agent further: L4 · Runs on its own against L1 · Needs a person.
- Find: Docs for agents: llms.txt, an OpenAPI and an MCP server; a docs page for each core task.
- Access: No key needed, sign-up in a browser.
- Use: 89% of the work.
- Pay: Payment not tried: robots.txt keeps our agent off the payment page for now.
- Leave: Leaves on its own: Nothing to cancel, made nothing to delete.
Through every checkpoint. Next: pay without a person
Runs on its own once it holds an existing identity: L4 · Runs on its own. Sign-up itself never got past a block on the signup page, so every later call ran on an existing account approved once through device-flow login. From there the work ran headless: REST, GraphQL, the gh CLI and the official MCP server created repos, commits, pull requests, reviews, issues and Actions runs, repeated across the core work. Codespaces refused our datacenter network, and the cloud coding agent needs a paid plan, so neither ran. Nothing to cancel on the free plan; account deletion is documented as web-only, so our agent revoked its grant and left the identity in place.
Gets furtherDriven 13 Oct 2026 · The drive and its evidence- Find: Docs for agents: llms.txt and an MCP server; a docs page for each core task.
- Access: Can't sign up.
- Use: Work not reached.
- Pay: Payment not reached.
- Leave: Leaving not reached.
Stopped at sign up: needs a human. Next: let an agent sign up on its own
Linear is a project-management and issue-tracking tool. An agent could not even create an account: the sign-up form shows a visible "prove you're human" check (Cloudflare Turnstile) right after the email is entered. So nothing behind sign-up (creating issues, projects, using the API, paying for a plan) could be tried.
Driven 16 Oct 2026 · The drive and its evidenceThe two summaries: written by our agents from each product's drive. How we write.
Where each one stops.
GitHubThrough every checkpoint
LinearStopped at sign up: needs a human
| Checkpoint | GitHub | Linear |
|---|---|---|
| Docs | Headless | Headless |
| Sign up | Not needed | Needs a human |
| Verify | Not needed | Not reached |
| Key | In a browser | Not reached |
| First call | Headless | Not reached |
| First action | Headless | Not reached |
| Pay | Not tried | Not reached |
| Cancel | Not needed | Not reached |
| Delete | Blocked | Not reached |
Headless or in a browser.
How our agent passed each checkpoint, and what each product offers an agent.
Checkpoints by route, API, MCP, CLI and SDK
| GitHub | Linear | |
|---|---|---|
| Headless (checkpoints) | 3 | 1 |
| In a browser (checkpoints) | 1 | 0 |
| Needs a human (checkpoints) | 0 | 1 |
| Blocked (checkpoints) | 1 | 0 |
| Not tested (checkpoints) | 1 | 7 |
| Work done | 89% of the core | Not measured |
| API | Yes | Yes |
| OpenAPI | Yes | No |
| MCP server | Yes (official) | Yes (official) |
| CLI | Yes | No |
| SDK | – | – |
| llms.txt | Yes | Yes |
Bot checks, and where a person was needed.
| GitHub | Linear | |
|---|---|---|
| Bot check | A challenge on the form stopped one way in (Email and password); it got in another way | Cloudflare Turnstile stopped it at sign up, with a visible challenge |
| Wall type | A challenge on the form | A check of our network and browser |
| First wall | None | Visible Turnstile on email sign-up |
| A person helped | No | No |
A challenge on the form: It asks every visitor to prove they are human. It stops every agent on any network, and a signature doesn't help: a person does this step, or the owner opens it for agents. A check of our network and browser: It judged where our agent connects from and its browser, not a person. An agent that works on its own from a data center doesn't get through unless the owner lets signed agents in. How agents in a person's own browser fare isn't measured yet. Where we drive from.
Level and part scores.
| GitHub | Linear | |
|---|---|---|
| Level | L4 · Runs on its own | L1 · Needs a person |
| Holds up | 59 of 100, fair | Not measured |
| Agent-ready setup | 100 of 100, good | 75 of 100, fair |
| Standards | MCP spec of 28 Jul 2026 (current) · MCP discovery (current) · OpenAPI 3.0.3 (2 versions behind) | MCP client registration (behind) |
Docs match is measured in an Audit. What the scores mean.
Pay, cancel and delete.
| GitHub | Linear | |
|---|---|---|
| Pay | Payment not tried | Payment not reached |
| Free plan or trial | Free plan | Free plan |
| Cancel | Not needed | Not reached |
| Delete the account | Blocked | Not reached |
How each one takes payment
- GitHub: Not tried yet: robots.txt keeps our agent off the payment page for now.
- Linear: Payment not reached.
GitHub vs Linear: the short answers.
Which can an AI agent use, GitHub or Linear?
GitHub gets an AI agent further: L4 · Runs on its own against L1 · Needs a person. With GitHub our agent went through every checkpoint; with Linear it was stopped at sign up (needs a human). Driven 13 Oct 2026 (GitHub) and 16 Oct 2026 (Linear).
Can an AI agent sign up for GitHub and Linear on its own?
GitHub: not needed. Linear: no, needs a human.
Do GitHub and Linear have an MCP server?
GitHub: yes (official). Linear: yes (official).
Can an AI agent delete its GitHub or Linear account?
GitHub: blocked. Linear: not reached.
The drives and the category.
Put your own shortlist side by side.
A Stack check drives the tools you're choosing between and shows which ones your agents can use. $490 for three tools, ready in 24 hours.