Worked example · 2026-06

Token of Power ($TOP) — low-float governance takeover mints 10B TOP in one block — Ethereum — 2026-06-09

Loss: ~$1.58M realised (944.2 WETH). Token of Power's governance ran on an Aragon DAO whose voting token (TOP) had a microscopic total supply of 16,384. The attacker acquired 8,192.000001 TOP — just over half the supply — which gave outright majority voting power, then exploited the absence of a timelock to propose, vote, and execute a malicious proposal in a single block. The proposal directed the DAO's TokenManager to mint 10,000,000,000 TOP directly to the attacker's contract; the attacker then dumped the minted TOP into the TOP/WETH Balancer V1 BPool, extracting 944.2 WETH (~$1.585M) and draining the pool to near-empty. The nominal 10B-TOP mint is bounded by the contract; the ~$1.58M realised is bounded by the Balancer pool's WETH depth — the standard mint-and-dump notional-vs-realised gap. The token's market value collapsed to near zero.
OAK Techniques observed: OAK-T9.003 (Governance Attack — the governance-contract design was the standing surface: an Aragon DAO with MiniMeToken voting, no timelock, no minting cap, and a quorum/pass threshold trivially clearable by a single majority holder, allowed propose → vote → execute in one block. The load-bearing flaw is the governance configuration, not a contract-code bug per se. See techniques/T9.003-governance-attack.md). OAK-T16.002 (Hostile-Vote Treasury Drain, direct-market-accumulation sub-shape — the attacker acquired voting power through legitimate on-chain acquisition of a thin token rather than via flash loan, held the majority position through the (single-transaction) proposal lifecycle, and voted a self-minting proposal through. This is the True Seigniorage Dollar / Build Finance DAO pattern restated at micro-float scale: when circulating supply is 16,384 tokens, "accumulate >50% of the governance token" costs little and needs no flash loan. See techniques/T16.002-hostile-vote-treasury-drain.md). OAK-T9.004 (Access-Control Misconfiguration — the executed proposal handed the attacker the TokenManager mint authority with no cap; an open mint reachable through a captured governance role. See techniques/T9.004-access-control-misconfiguration.md). OAK-T5.001 (Hard-LP-Drain — the minted TOP was swapped against the TOP/WETH Balancer V1 pool for 944.2 WETH, draining the pool. See techniques/T5.001-hard-lp-drain.md). OAK-T7.001 (Mixer-Routed Hop — the attacker's address was funded by a ~$2.7M ETH withdrawal from Tornado Cash and proceeds were routed back through Tornado Cash, both ends of the operation mixed. See techniques/T7.001-mixer-routed-hop.md).
Attribution: pseudonymous. TRM Labs and PeckShield analysed the incident; the attacker address was funded from Tornado Cash and proceeds were laundered back through it, leaving no named individual or group. The exploit transaction(s) executed on Ethereum on 2026-06-09.
Key teaching point: Governance security degrades with float. Every mitigation that defends a DAO against a flash-loan vote takeover (snapshot-based voting power, multi-block windows) is irrelevant if the governance token's total supply is small enough to simply buy a majority outright — and then the only remaining defence is a timelock + minting cap, both of which Token of Power lacked. With 16,384 TOP in existence, acquiring 8,192 was cheap and required no flash loan, so the attack collapsed to the T16.002 direct-accumulation shape; with no timelock, the attacker compressed propose-vote-execute into one block, achieving flash-loan-style atomicity without a flash loan. The defender lessons are concrete: (a) a timelock with a public review window on any treasury- or supply-affecting proposal is the single highest-leverage control for low-cap DAOs, because it is the one mitigation that does not depend on voting-power distribution; (b) mint authority must carry a hard cap independent of governance, so capturing a governance role cannot become an unbounded mint; and (c) circulating-supply-vs-quorum ratio is a measurable governance-risk metric — a token whose entire supply is cheaply acquirable is a standing governance-takeover surface regardless of flash-loan resistance.

Summary

Token of Power (TOP) is a small-cap Ethereum project whose on-chain governance was implemented as an Aragon DAO using MiniMeToken-based voting. The TOP governance token had a total supply of only 16,384, and the DAO configuration had no timelock between proposal passage and execution, no minting cap, and voting thresholds a single majority holder could clear.

On 2026-06-09, an attacker — funded by a ~$2.7M ETH withdrawal from Tornado Cash — acquired 8,192.000001 TOP, just over half the supply, obtaining outright majority voting power. In a single transaction, the attacker created, voted on, and executed a malicious governance proposal that instructed the DAO's TokenManager to mint 10,000,000,000 new TOP to an attacker-controlled contract. The attacker then sold the minted TOP into the TOP/WETH Balancer V1 BPool, extracting 944.2 WETH (~$1.585M) and draining the pool's liquidity to near zero. Proceeds were routed back through Tornado Cash. TOP's market value collapsed and trading activity ceased. TRM Labs and PeckShield attributed the loss to the Aragon DAO configuration — specifically the absence of a timelock combined with the microscopic token supply and uncapped mint.

Timeline (UTC)

When	Event	OAK ref
Pre-2026-06-09	TOP governance = Aragon DAO + MiniMeToken; total supply 16,384; no timelock, no mint cap; TokenManager mint reachable via passed proposal	(standing T9.003 surface)
2026-06-09	Attacker funds operating address with ~$2.7M ETH from Tornado Cash	T7.001 (funding)
2026-06-09	Attacker acquires 8,192.000001 TOP (>50% of supply) → majority voting power	T16.002 accumulation
2026-06-09	In a single block, attacker proposes + votes + executes a malicious proposal; TokenManager mints 10,000,000,000 TOP to attacker contract	T9.003 + T9.004 execution
2026-06-09	Attacker swaps minted TOP into the TOP/WETH Balancer V1 pool for 944.2 WETH (~$1.585M); pool drained	T5.001 extraction
2026-06-09	Proceeds routed back through Tornado Cash; TOP value collapses	T7.001 (laundering)

What defenders observed

Pre-event (governance-design layer). Circulating-supply-vs-quorum ratio was catastrophically favourable to a single acquirer (16,384-token supply), there was no timelock, and mint authority was uncapped. All three are pre-deployment governance-design signals; OAK-M17 (time-locked governance) and OAK-M16 (governance-contract audit) are the relevant mitigations, and a circulating-supply-vs-quorum-threshold check is the leading pre-event indicator for the low-float sub-shape.
At-event (single-block propose-vote-execute). A proposal that is created, voted, and executed in the same transaction, by an address holding a majority of a thin governance token, with an execution payload that mints to the proposer, is the direct exploitation signature. Proposal-payload static analysis (OAK-M02) — alerting on payloads that mint privileged supply or direct value to the proposing address — is the detection control, but with no timelock there is no review window in which to act.
At-event (mint + dump). A 10B-TOP mint (orders of magnitude over the 16,384 supply) followed immediately by a Balancer-pool swap is the mint-and-dump fingerprint; a supply-magnitude monitor would fire on the mint, and the realised loss is set by Balancer pool depth (944.2 WETH).
Post-event (mixer on both ends). The operating address was Tornado-funded and proceeds were Tornado-laundered, the standard privacy-routed cash-out; the both-ends-mixed pattern is itself an attribution-frustration signal.

Public references

[bitgettop2026] — Bitget News, "Attacker cleans out $1.6M from Token of Power (TOP) in Aragon DAO exploit" (Aragon + MiniMeToken, 16,384 supply, >50% acquisition, single-tx propose/vote/execute, TokenManager mint of 10B TOP, 944.2 WETH from TOP/WETH Balancer V1, Tornado funding/laundering): https://www.bitget.com/amp/news/detail/12560605451917
[cryptonewstop2026] — crypto.news, "Token of Power exploit drains $1.58M from Balancer pool": https://crypto.news/token-of-power-exploit-drains-1-58m-from-balancer-pool/
[ambcryptotop2026] — AMBCrypto, "Governance takeover lets attacker mint 10B TOP tokens in $1.5m exploit": https://ambcrypto.com/governance-takeover-lets-attacker-mint-10b-top-tokens-in-1-5m-exploit/
[bitcoinisttop2026] — Bitcoinist, "Token Of Power Governance Exploit Drains $1.58 Million In WETH, TRM Says" (TRM analysis: no-timelock propose-vote-execute in one block): https://bitcoinist.com/token-of-power-governance-exploit-drains-1-58-million-in-weth-trm-says/
[cryptotimestop2026] — Crypto Times, 2026-06-10, "One Vote, $1.58M Gone: TOP Token Hit by Alleged Governance Attack": https://www.cryptotimes.io/2026/06/10/one-vote-1-58m-gone-top-token-hit-by-alleged-governance-attack/
[crowdfundinsidertop2026] — Crowdfund Insider, 2026-06, "Tornado Cash On Ethereum: $2.7M ETH Withdrawal Used In TOP Governance Takeover": https://www.crowdfundinsider.com/2026/06/285684-tornado-cash-on-ethereum-2-7m-eth-withdrawal-used-in-top-governance-takeover/

Discussion

Token of Power is the micro-float restatement of OAK's hostile-vote-mint pattern, joining True Seigniorage Dollar (BSC, 2021-03) and Build Finance DAO (Ethereum, 2022-02) as worked examples of T16.002's direct-market-accumulation sub-shape. Its instructional value is the explicit demonstration that flash-loan-resistance does not defend a thin token: the post-Beanstalk industry mitigation (snapshot-based voting power) is aimed at attackers who hold the position only inside a flash-loan block, but when the entire supply is 16,384 tokens the attacker simply buys a majority and holds it through every snapshot — exactly the T16.002 case that snapshotting leaves open. The case therefore reinforces the T16.002 thesis that the load-bearing mitigations live above the voting-power-computation layer: a timelock review window and a governance-independent mint cap.

The single-block propose-vote-execute is also a useful boundary marker against T16.001. Operationally the attack looks atomic, like a flash-loan governance takeover — but there is no flash loan; the atomicity comes from the absence of a timelock, not from same-block borrow-and-repay. OAK maps it as T9.003 (governance-contract configuration) + T16.002 (direct accumulation), not T16.001, precisely because the acquisition mechanism is outright purchase of a low-float token rather than flash-borrowed voting weight. Contributors writing future low-cap-DAO governance examples should record the circulating-supply-vs-quorum ratio and the timelock presence/duration as the two decisive calibration variables, because together they determine whether a governance token is a standing takeover surface.