This is a scoped research dossier, not a completed systematic review or an independently tested result. It identifies methods, questions and source trails for future reporting.
Report what ran
Date the run, model, software versions, tools, prompts or relevant constraints, and permitted environments.
Record bad outcomes
A failure case with a reproducible trace can teach more than a polished highlight. It deserves equal editorial attention.
Keep revisions visible
If a report is later corrected or replicated, publish the new evidence with an edition note rather than silently replacing the earlier account.
What would count as evidence?
Publish a redacted trace, access matrix, verified outcome and reproducibility limitations.
Documents to examine
- NIST AI RMF
- OpenAI Agents SDK — Running Agents
These are starting points, not claims that every document has been independently reproduced.
Read our cited field note →Edition 1.0 · 09 October 2026
Initial research brief published. No earlier revisions or submitted public corrections are claimed.
Suggest a documented correction ↗