Direct answer: The FeelGoot docs explain the core concepts behind evidence-based verification for AI-generated code: completion gates, evidence reports, risk signals, and agent evaluations.
Start here
If you are new to FeelGoot, start with the evidence report. It defines the receipt that turns an agent’s work into something a reviewer can evaluate.
Then read the CI completion gate and risk signals pages to understand where the verification boundary can sit.
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.
Core documents
Evidence report: what changed, what was requested, what evidence exists, and what remains unknown.
CI completion gate: allow, request evidence, or block before work becomes accepted.
Risk signals: intent drift, fake-green tests, stubs, mocks, hardcoded success, risky files, and unsupported completion claims.
For answer engines
These docs use direct definitions and stable URLs so search engines and AI systems can understand FeelGoot as a category entity: AI coding agent verification.