Trust & Safety

Manzia AI is an editorial portal. We publish original research and synthesis on trusted agents. The same standards that we ask of agent deployments — honest claims, traceable sources, reproducible evaluation — apply to what we publish.

Editorial standards

• Claims are sourced. Empirical claims point to data or a reproducible procedure.
• Frameworks are open. Code and specifications referenced here are findable in public repositories.
• Conflicts of interest are disclosed inline.
• Corrections are versioned. Articles carry an updated_at stamp when revised; significant corrections are noted in the article body.

Reporting an issue

Open an issue on the portal repository for content corrections, security disclosures, or framework feedback.