Evidence first, prose second, schedule a distant third.

Every article goes through editorial review, technical review, privacy review, and final human approval before publication. If a claim is weak, the fix is stronger evidence or a smaller claim. It is not a brisker headline.

Publication rules

• Claims trace back to evidence.
• Assumptions are labelled as assumptions.
• Recommendations stay proportionate to whatever actually supports them.
• Tradeoffs sit next to the advice they qualify.
• Private infrastructure detail does not survive into public work.

Recommendation rule

Product mentions are allowed when they help explain a real build, measurement, limitation, or decision. Broad "best" claims are not. Neither is borrowed authority from other people's test benches dressed up as first-hand judgement.

What fails review

• Invented measurements or invented product testing.
• Broad “best” claims without proof.
• Borrowed bench results dressed up as first-hand judgement.

Final approval stays with the human editor. The workflow can be automated. Accountability cannot.