Air traffic control for AI coding

nool debug blast-radius

Blast-radius analysis shows which system boundaries a change may affect so teams can catch AI agent scope drift before acceptance.

  • Analyze files after implementation and before proposal.
  • Use risk scores to decide when to split work or add justification.
  • Connect unexpected impact to tests, findings, or human review.
nool debug blast-radius src/pages/Setup.tsx --compact

Why it matters

Blast radius turns a file list into an impact question: what system boundaries could this change affect, and did the agent touch more of the system than the task allowed?

Analyze files

Pass one or more paths after implementation and before proposal. Use focused paths instead of the whole repository when you already know the edited surface.

nool debug blast-radius src/pages/Setup.tsx src/data/blogs.ts --compact

Risk interpretation

Low risk usually means the change stayed inside one local feature. Medium risk calls for a clear proposal intent and tests. High risk is not an automatic rejection, but it should trigger stronger justification, reviewer attention, or a split into smaller tasks.

  • Low: expected docs, test, or local component edits.
  • Medium: shared component, routing, pricing, or data files.
  • High: policy, security, persistence, build, release, or cross-module behavior.

Failure mode it prevents

Git can show that a file changed. It cannot tell you that a checkout task unexpectedly altered license enforcement language or that a docs update changed a shared navigation component used across every page.