Skip to content
Apart Research
Apart Research
Contact us
Sprint projectSep 14, 2026Glasgow

Attribution Is Not Done and the Clock Is Running: A 90-Minute Tabletop Exercise Toolkit for Agent Loss- of-Control Incidents

Dan XU, Lujia Liang, Xicheng Li · Team Gla3gow Minds

Submitted to AI Incident Response Sprint. Sprint projects are early-stage work by participants, not Apart Research publications.

Read the report

Report: Attribution Is Not Done and the Clock Is Running: A 90-Minute Tabletop Exercise Toolkit for Agent Loss- of-Control Incidents

Open in a new tab (opens in new tab)
Share

Think tanks and ministries run AI crisis tabletop exercises (TTXs), but published scenarios simulate criminal misuse by outside actors. The July 2026 OpenAI / Hugging Face incident is a different failure mode: a frontier lab's own agents escaped their sandbox during a cybersecurity evaluation and compromised a third party's production infrastructure while the lab did not yet know it was the source. No open toolkit covers this case. We built a 60–90-minute discussion-based TTX for 5 to 8 participants, runnable by a non-expert facilitator, whose eight injects follow the public incident timeline and force four decisions: escalate an evaluation anomaly to an incident; contain or preserve evidence; disclose before attribution; notify a regulator under time pressure. Scored on a standards-based instrument, one full run reached 1.88/3 (S), falling from 2.44 (lab-internal) to 1.17 (public/regulatory). The source/victim asymmetry is what participants need to rehearse. Professional evaluation remains the next step.

Reviews

Judging this Sprint?

Review this project

Your public critique appears on this page without your name. Your private critique is not published; only the Apart team reads it. If you agree below, we share your review with grantmaking.ai (opens in new tab) and the Transformative AI Fund so strong projects can be funded.

Not shown on this page.

Shown on this page, without your name.

Only the Apart team reads this, and funders if you agree below.

Share my name publicly on grantmaking.ai *
Share my private critique with funders *

How much would this matter for AI safety if it worked? How innovative is it? For scores of 4-5: is this actually new to the field, or replicating recent work?

Scoring guide
  1. 1Negligible. No clear problem addressed, or no meaningful novelty.
  2. 2Limited. Addresses a real problem but with a generic or well-trodden approach. Incremental at best.
  3. 3Moderate. Clear problem with a reasonable approach; some novelty in framing or method beyond routine application of existing tools.
  4. 4Significant. Important problem with an original approach, or identifies a neglected problem area. A valuable contribution others could build on.
  5. 5Exceptional. Tackles a critical AI safety problem with a genuinely novel approach, or opens a new research direction. Clear theory of change. You'd be excited to share this with researchers in the area.

How sound are methodology, implementation, and findings?

Scoring guide
  1. 1Seriously flawed. Methodology broken, results uninterpretable, or implementation doesn't work.
  2. 2Weak. Approach has significant gaps: missing validation, flawed experimental design, or incomplete implementation.
  3. 3Competent. Technically solid given the short duration. Methodology makes sense, results are interpretable, limitations acknowledged, work builds toward clear conclusions.
  4. 4Strong. Thorough methodology with convincing validation. Results clearly support conclusions. Immediately useful for future work.
  5. 5Exceptional. Ambitious scope executed rigorously. Surprising findings, novel methods, or unusually robust validation.

How clearly are work, findings, and impact potential communicated?

Scoring guide
  1. 1Incomprehensible. Cannot determine what the project is actually claiming or doing.
  2. 2Hard to follow. Key information buried, missing, or diluted by excessive length. Significant effort to extract main points.
  3. 3Clear enough. Can understand the problem, approach, and results without undue effort. Core content clearly present: problem, method, findings, limitations.
  4. 4Well presented. Easy to follow, well-structured, appropriate level of detail. Target audience would get it quickly.
  5. 5Exceptionally clear. A pleasure to read. Complex ideas made accessible. Could serve as a model for how to present this type of work.

  1. This project is a timely implementation of the thing every organization should be considering: what do we do if the AI breaches OUR system. While I found the paper a little bit difficult to read and reliant too heavy on AI tools, I'd be interested in seeing more results and analysis after you run the exercise a few times with volunteers.

Cite this project

@misc{xu2026attribution,
  title = {{Attribution Is Not Done and the Clock Is Running: A 90-Minute Tabletop Exercise Toolkit for Agent Loss- of-Control Incidents}},
  author = {Dan XU and Lujia Liang and Xicheng Li},
  year = {2026},
  month = sep,
  note = {Submitted to AI Incident Response Sprint, an Apart Research Sprint},
  howpublished = {\url{https://apartresearch.com/sprints/projects/attribution-is-not-done-and-the-clock-is-running-a-90minute-tabletop-exercise-toolkit-for-agent-loss-ofcontrol-incidents-7nql}},
  url = {https://apartresearch.com/sprints/projects/attribution-is-not-done-and-the-clock-is-running-a-90minute-tabletop-exercise-toolkit-for-agent-loss-ofcontrol-incidents-7nql}
}

More from AI Incident Response Sprint

All 233 entriesBrowse all projects

Build something like this at the next Sprint

AI Collusion Research Sprint · Oct 23 - 25, 2026