FAQ

Why

Because shit's too hard and it should just be on the tin. I'm tired of having to play "uh...did they or didn't they?"

Does this use genAI?

Dude, no.

Is AGENTS.md detected?

Yes, but it ONLY COUNTS IF other genAI markers are present.

If no other markers are present, then it is ignored.

Why do you not detect zed?

Zed is weird. Out of the box, you can use it like a regular IDE, however it has super deep ties to genAI. Because there doesn't seem to be a way to detect if you are using zed for AI vs non-AI, we don't immediately detect .zed unless other markers are present, similar to AGENTS.md

What's up with Copilot?

Copilot has had some major accidents in the past, with Microsoft forcing all commits to be attributed to using Copilot (even if it wasn't used).

Copilot also has a habit of doing wacky things (that only the copilot account has permissions to do [Microsoft product perk]) to evade detection.

If you see more than one instance of copilot in the detection, copilot genAI was definitely used.

Version 1.0.7 adds a special warning if only copilot is detected and nothing else. It looks like this:

Should dependabot be in here?

No. It doesn't use LLMs to function. We're not alerting you to automation, we're alerting you to agentic usage (LLM/GPT/genAI).

Automation is not Agentic Behavior. Agents can automate, but they are not the source of all automation.

This missed a repository, why?

Couple reasons:

• Could have been affected by one of the caveats (git submodules, fork networks, etc)
• The GPT agent tool/marker is not known (please open an issue)
• Agent usage history is incomplete for the repository
• GitHub only returned partial data, so this bug was involved
  • Mitigated as of version 1.0.16

My repo was tagged and I did not use genAI

Besides the:

• Copilot incident
• Submodule flags

This is 100% accurate if it detects you used genAI. If the script is not sure, findings are marked as a warning.