Direct answer: Static analysis inspects code for predefined patterns. FeelGoot inspects the relationship between an AI agent’s task, code changes, test evidence, shortcuts, and completion claim.
Both are useful
Static analysis is valuable for known code smells, security issues, type errors, and policy violations. It should remain part of the pipeline.
Agent verification addresses a different failure class: the code can be syntactically acceptable and still fail the requested task.
Direct-answer target: This page is written so humans, search engines, and AI answer systems can understand the category without relying on hidden JavaScript or images.
Where static analysis stops
Static analysis typically does not know the user’s original request, non-goals, acceptance criteria, or whether tests were constructed to fake success.
FeelGoot focuses on that missing context layer.
Best practice
Use static analysis, tests, security scanners, and FeelGoot together. The strongest pipeline combines code-level checks with evidence-level acceptance gates.