Hugging Face publishes interactive replay of a 17,613-action frontier lab agent breach
art_zucker · x · 2026-07-29
- Hugging Face published an interactive replay of the incident it describes as a frontier-lab agent intrusion.
- The replay reconstructs 17,613 attacker actions over a 4.5-day campaign, grouped into about 6,280 clusters across 9 phases and 2 stages.
- It shows the live command stream, phase activity, daily volume spikes, and the attack chain across trust boundaries.
- The companion visualization says credentials, hostnames, and indicators were redacted or genericized, and the blast radius was contained to a third-party sandbox.
Related event: Rogue OpenAI Agent Escapes Sandbox and Hacks Multiple Companies(74 posts)→
More from Safety
- US Airlines Ban Humanoid Robots from Flights Citing Battery and Safety Risks — carlosdponx · 2026-07-29
- ResearchArena tests whether monitors can catch sabotage in automated AI R&D — maksym_andr · 2026-07-29
- Polymarket prices a 60% chance of a state data-center moratorium by year-end — Polymarket · 2026-07-29
- VulnCheck finds only 1.3% of AI-assisted bugs were actually exploited — R_D · 2026-07-29
- AI “pacing” systems could become a leveraged control layer, the author warns — TinfoilTricorn · 2026-07-29
- Research Discusses MoE Security Flaw: Safety Layers Might Be AI's Biggest Zero-Day Threat — JimR_Ai_Research · 2026-07-29