Editorial Policy

Independence

Our checks and scoring methodology are determined independently of commercial relationships. Sponsors do not influence scoring rules or recommendations.

Selection criteria for checks

We prioritize checks that are observable over HTTP, widely referenced by agent ecosystems, and actionable for engineering teams. We avoid opaque heuristics that cannot be validated.

Originality standard

We maintain original editorial copy for product pages, policy pages, and prompt content. We do not publish scraped third-party documentation as primary page content.

Scoring methodology

For the content production profile, checks are evaluated as pass, fail, or neutral. Neutral checks are informational and excluded from score denominator. Reported score is passed/scored checks.

Accuracy and false positives

We run best-effort HTTP checks. Some hosting setups can cause false negatives. If you think your result is inaccurate, please contact us.

Prompt generation policy

Fix prompts are designed as implementation starting points. They are intentionally concise, not full deployment runbooks, and should be adapted to your stack, threat model, and change controls.

Versioning and updates

We revise checks as standards and platform behavior evolve. Significant methodology changes are reflected in this policy and in public-facing documentation.