Compound vote approved 499,000-COMP goldCOMP allocation amid governance alarm

Compound governance approved Proposal 289 on July 28, 2024, authorizing a one-year allocation of 499,000 COMP from the lending protocol’s treasury to a goldCOMP investment arrangement associated with a group called the Golden Boys.

The proposal received 682,191 votes in favor and 633,636 against, according to contemporaneous reporting, a margin of 48,555 COMP votes. Compound’s governance interface records the corresponding rounded totals as approximately 682,190 for and 633,640 against, with the affirmative vote exceeding the 400,000-COMP quorum requirement.

The result mattered beyond the allocation itself. It demonstrated how a concentrated voting bloc could overcome opposition from prominent Compound delegates while operating within the protocol’s token-weighted rules. Critics called the episode a possible governance attack; the proposer rejected that characterization. On July 28, the competing claims remained allegations and interpretations rather than an adjudicated finding of theft, fraud or technical exploitation.

What Proposal 289 authorized

The proposal described the allocation as 5% of Compound’s non-interest-bearing COMP treasury holdings. Its two specified actions were to transfer 499,000 COMP to a deployed Trust Setup contract and place that contract in a phase allowing the Golden Boys multisignature wallet to initiate an investment.

Under the proposer’s design, treasury COMP would enter the goldCOMP vault, which issued a wrapped representation of deposited COMP. That token could be paired in a 99/1 Balancer pool as part of the intended yield strategy. The proposal stated that rewards would eventually be converted through a defined process and returned to Compound’s Comptroller.

The vote did not itself complete the transfer. Passing Compound governance created authority for the listed calls to proceed through the protocol’s queue and timelock sequence. Consequently, the event-day record establishes approval, not that 499,000 COMP had already left the treasury or produced any return.

Why delegates objected

Wintermute Governance had argued in Compound’s forum that the DAO could change the Trust Setup’s phase but could not independently complete a withdrawal. A divestment would still require the Golden Boys multisignature wallet to call the relevant functions. Wintermute also warned that COMP placed in the goldCOMP vault would delegate its voting power to that multisignature wallet.

That second point made the proposal a governance-security question as well as a treasury decision. If the allocated COMP retained usable voting power under the recipient’s control, the transaction could strengthen the same bloc that had advanced the proposal. The proposal’s supporters presented the arrangement as a constrained investment structure; opponents questioned whether its constraints supplied an effective clawback mechanism.

OpenZeppelin’s Compound security adviser had raised concerns beginning in May 2024 after observing five accounts withdraw and delegate a combined 230,333 COMP in similar patterns. The adviser explicitly treated common control as an assumption based on those patterns, not a proven identity link. The forum notice nevertheless urged delegates to scrutinize the resulting concentration relative to Compound’s 400,000-COMP quorum threshold.

The third attempt changed the stakes

An earlier proposal, numbered 247, sought a related treasury allocation but was canceled after objections. Proposal 279 subsequently used the Trust Setup structure and sought 92,000 COMP but failed. Proposal 289 retained the structure while increasing the requested allocation to 499,000 COMP.

Before the final vote, the proposer said the Trust Setup did not permit funds to be stolen or diverted. OpenZeppelin’s adviser expressed the opposite assessment as a personal opinion and urged cancellation. The narrow July 28 result settled the formal vote, not that factual and institutional dispute.

What remained unknown on July 28

No event-day record established that the strategy would generate its advertised yield, that the DAO could recover the allocation without cooperation, or that the approved calls would ultimately execute. This reconstruction makes no COMP price, return, trading-volume or total-value-locked claim because the governance result alone cannot isolate a market effect.

The verified significance was therefore narrower: a valid Compound vote approved a large treasury allocation despite organized security objections, revealing the tension between permissionless token voting and institutional confidence in a protocol’s governance safeguards.