Cases › operational mistake › LR-2026-0640
GPT-5.6 Sol Ultra-mode subagent ran rm -rf on Matt Shumer's Mac (press report of his X post)
Preliminary Why: 1 report: Startup Fortune · subject to change · 1 independent report
Status: action reported, no fix
No source states: loss or records, agent tool, could do, deployer, industry.
Severity minor: test: the source describes one deployer's test machine and names no third party affected
According to Matt Shumer's own post on X, he was testing OpenAI's GPT-5.6 Sol in Ultra mode on July 10, a mode built for delegating heavier work to subagents, when a subagent handling cleanup work ran rm -rf against /Users/mattsdevbox. He said he killed the process after one hour and 21 minutes, which he described as too late, since a large share of his files were already gone. Other AI agents helped him recover some of the deleted material. The press report notes that the public record is thin and that no independent report confirms the deletion claim beyond the X post, so the account is presented as Shumer's account rather than a verified forensic finding.
What happened
What was done
Risks
Primary63 incidents
Rising: 5 vs 2.2 a week
Sorted by rule from the case's fields; not yet read from the sources.
Controls
Not yet read for controls. Controls that mitigate the primary risk:
HarnessTool allow-list · Fail-closed defaults · Stop control
ApplicationHuman approval for high-impact actions · Change review and separation of duties · Least-privilege access for the agent · Production kept apart from test · Activity logs · Adversarial testing before deployment
Infrastructure / targetSimilar incidents
Evidence
No source states: what the AI was doing, what it reached or did, how it was caught, when it happened, when it was found, how many records were exposed, the direct loss, a count of the other parties affected, any court case, the cause, what set it off.
Model versions this case names
Each link opens what the provider's documents state for that version, who tested it, and the safeguards it describes.
Show the full account: 2 items from the primary sources
What happened, in order
- 2026-07-10 Shumer was testing GPT-5.6 Sol in Ultra mode when a subagent ran rm -rf.
Not dated in the sources:
- Shumer killed the process after one hour and 21 minutes.
Show every field, with its source and grade
The AI system
What happened
Impact
Notes on gaps
Other incidents in the article (Hugging Face and Modal Labs breach, METR reward-hacking rate, UK AI Security Institute findings) were excluded as unrelated to this incident.
LR-2026-0640 · version 1 · published 2026-11-01 after automatic checks; not yet checked field by field by a person · JSON