Frequently Asked Questions

What does AGI-HIVE do?

It ingests your repo, runs Sentinel scans + Hunger tasks, routes fixes with P-Axis, and logs evidence so you can ship safer patches faster.

How is this different from linting?

Linting catches syntax/style. AGI-HIVE prioritizes vulnerabilities, gaps, and improvements, then routes and previews fixes with risk and evidence.

What languages are best supported?

We focus on TypeScript/JavaScript first. Sentinel/Hunger heuristics work on other languages for common patterns (security strings, TODOs, gaps).

How long does analysis take?

Small repos (≤100 files) typically scan in ~30-90 seconds depending on network and GitHub fetch time.

What engines are involved in a fix?

Forge orchestrates Sentinel (scan), Hunger (tasks), Worldview (map), P-Axis (routing), Evidence/Narrator (logging & explanation), and Cortex (learning).

How are patches verified?

You review the diff, can run risk checks, and every action is logged to Evidence. Human approval stays in the loop.

Can I call this from CI/CD?

Yes. Use the API routes (Sentinel scan, Hunger tasks, P-Axis route, Evidence append) to automate checks before merge.

What about privacy?

Code stays within your session. Evidence logs store metadata about actions, not full source content. Self-hosting and private provider keys are supported.