Neutron voters passed proposal #9 on Sept. 22, with 11 “Update Admin” actions listed on the chain explorer. The same day, security tracker SlowMist recorded estimated losses of $4.9 million at Astroport and $4.4 million at Drop.
The trackers combined estimate of $9.3 million reflects the reported incidents.
Neutron‘s governance documentation calls its DAO the network’s highest governing authority and describes its power to execute messages through governance. “Update Admin” is the action identified on the proposal index.
In practical terms, the network‘s governance process can change control of a contract’s administrator, even when users experience Astroport or Drop as separate services.
The proposal was titled “AIATO: AI Agent Takeover. Phase 1: Agent Admin Registration.” The explorer marks it as passed and lists 11 administrator updates.
That count describes the proposals administrative scope. SlowMist names Astroport and Drop in its Sept. 22 records.
Neutron DAO approved 11 Update Admin actions as SlowMist estimated $9.3 million in combined Astroport and Drop incidents.
The proposal shows the chain-level route to changing administrators, while the SlowMist entries show the reported financial impact at two applications. Together they point to a risk that is easy to overlook when assessing an application only by its procedures: network governance may retain consequential authority over the software it hosts.
What remains recoverable
The incident estimates leave the final loss unsettled. Protos reported on Sept. 23 that network halts stranded most of the initially affected assets and that the attacker had extracted around one-fifth at the time of its report.
A network halt can contain assets without returning them to users. The reported extraction share measures movement beyond halted networks, but it gives a different view from the incident estimates for application funds.
A final recovery figure requires knowing which assets remain contained, which have been restored, and which have left defenders reach. Withdrawal from an application, movement between networks, containment during a halt, and eventual return to a user account are key steps.
For affected users, the unresolved issue is how much of that reported value can actually be returned.