---
name: twofold-two-3
description: Continuation of [[twofold-two-2]] from 18 Sep 2026, after that file neared its cap. The three-iteration re-audit Ben demanded, which overturned four of my own claims from the same day - the second contract stack, TwoStakingUSDG (Vault 2), the real reward spend, and the operator net-buying. Read this BEFORE twofold-two or twofold-two-2, because it corrects both.
sources: [cowork]
aliases: [two-3, twofold-3, twofold part 3]
---

# TWO / Twofold part 3. 18 Sep 2026, three-iteration re-audit.
Ben: "Double-check all your information... do three iterations." He was right to ask. **Four claims I made earlier the same day were wrong.** Corrections first, at the top, naming the number that misled.

## CORRECTION 1: "staking collapsed to 5.47%" was WRONG. There is a Vault 2 I never looked for.
The @twofoldfi tweet said "vault 1 ... now retired" and I logged that without finding vault 2. It is in their own `/docs` page, which lists every contract.
**TwoStakingUSDG = `0x9CF18bB1dD9AfBF75B579Cc0C473B2975c16E9e3`**, 4,103 bytes. Docs: *"Vault 2: stake TWO, earn USDG; per-second stream, 7-day unstake cooldown, 400,000,000 TWO cap."* Verified on chain: `rewardToken()` = USDG `0x5fc5360d0400a0fd4f2af552add042d716f1d168`, `cap()` = 400,000,000 TWO, `cooldown()` = 604800 = **7 days** (Vault 1 was 1 hour).
- **`totalStaked()` = 116,538,217 TWO.**
- Vault 1 `0x06E463fDa4BEb4aA096142E673240aB9719fB3A9` = 54,728,708.
- **TRUE TOTAL STAKED = 171,266,925 TWO = 17.13% of supply**, not the 5.47% I reported. Staking did not collapse, it **migrated**, into a vault with a 7x longer cooldown. That makes the float STICKIER, not looser.
- Unresolved: Vault 2's TWO balance is 178,246,878 against `totalStaked()` 116,538,217, a **gap of 61,708,661 TWO (about $12,300)**. Could be the 7-day unstake queue (i.e. supply already heading out) or a reserve. `pending(address)` exists but no aggregate getter; Blockscout is Cloudflare-blocked from the container so the ABI could not be pulled. **Open item.**

## CORRECTION 2: `0xd1BcbCCa41f3bdb6b4812652959c6dF725ea2Ac0` is the NEW DualPoolHook, not a "JIT routing contract".
Their docs: *"DualPoolHook, Uniswap's audited v4 hook, deployed unmodified; every current pool (since 2026-09-05)."* My label was wrong; the mechanism description (withdraw, fill, return) was right, because that is what the hook does. The $14,159,859.20 of USDG cycling through it in 7 days is the hook's JIT plumbing, and per the standing rule it is **not** volume.

## CORRECTION 3: "the founder pays $200/day in rewards" was WRONG in both directions.
`dailyUsdg: 200` is a **ceiling**, not the spend. The real deposit-reward accrual is capped by `maxAprBps` 5000 (50% APR) applied to `weightUsd`:
`ratePerSec` 7221122685185185185185185182 / 1e27 = 7.2211/sec x 86400 = 623,905 = `dayBudget`, in USDG base units = **$0.62/day** of deposit rewards. Only 6 of 32 weighted pools carry any weight at all.
**But the STAKING stream is separate and much bigger, and I never measured it.** 14-day sweep of USDG in and out of TwoStakingUSDG:
- **Paid out to stakers: $4,432.33 in 246 legs to 68 wallets = $316.59/day.** First payout 4 Sep 09:52 BKK, still running 18 Sep 08:41 BKK.
- **Funded: $6,355.00, 952 legs of about $5 each, 100% from the operator EOA `0xC8720447712e6C4c851B3884b4Ec93F9cE8aD5fD`. Nothing from the harvest path, nothing from fees.**

## CORRECTION 4: "the operator sold zero" was too narrow, and the direction is the opposite of what I implied.
Over **14 days** the operator moved 22,103,317 TWO in and 11,390,483 out. **Net +10,712,834 TWO. He is accumulating.**
7,495,572 of the outbound went to the PoolManager. **In Uniswap v4 a transfer to the PoolManager is a sell OR a liquidity add, and the two are indistinguishable at the Transfer level.** For an operator who seeds his own pools it is most likely LP. **I cannot call it a sale and did not.** This is a real hole in the seller-ledger method on v4 and belongs in onchain-forensics: *for an address known to provide liquidity, a PoolManager transfer is not evidence of selling; resolve it against the v4 Swap and ModifyLiquidity events before labelling.*

## What the three passes CONFIRMED (unchanged)
- **TVL is genuinely tiny: $28,771** across both stacks (USDG leg x2, per pool, from their `/api/pools` `reserves`). Hook 1 (retired, withdraw-only) 30 pools / $3,335. Hook 2 (current) 34 pools / $25,436.
- **Swap fees $11.17/day** across 64 pools and 1,472 swaps in 24h, at each pool's real fee tier. Hook 2 does 1,446 of those swaps and 1,151 of them are one WETH/USDG pool.
- **The value-accrual path the docs promise is functionally dead.** OperatorControllerV2 `0x705F4f94344feD3f7BA98300F8bC04fdAE570340` "harvests gain above basis to the staking vault": **it has moved $0.17 in 14 days**, in and out, in dust legs.
- **Zero tier-one coverage, re-verified.** theunipcs, pingucharts, CrashiusClay69, 0xkioto all return 0. blknoiz06 returned 5 hits which are all **the English word "twofold"** in unrelated posts (Solana revenue, cycle thesis). Do not record that as Ansem coverage.
- LP still locked, NFT 1076283, liquidity byte-identical.

## THE NUMBER THAT MATTERS MOST, and it is new: the founder's total burn rate
Every USDG leg out of the operator EOA, 14 days:
| destination | $ | what |
|---|---|---|
| `0xd1bcbcca41f3bdb6b4812652959c6df725ea2ac0` | 17,831.68 (46 legs) | seeding the NEW hook's pools |
| `0x9cf18bb1dd9afbf75b579cc0c473b2975c16e9e3` | 6,355.00 (952 legs) | staking rewards |
| `0x8366a39cc670b4001a1121b8f6a443a643e40951` | 5,153.79 (32 legs) | PoolManager, LP or buys |
| `0xf64c3dfb1f246555d4a9597fb4ecddb3a8e59a07` | 2,565.00 (2 legs) | unlabelled EOA, not identified |
| `0x127b3f3b7769f659c5edbff8b4005443f19faac0` | 500.00 (5 legs) | seeding the old hook |
| `0xd5e88a436f5068697f27b18cd55931462126e3ff` | 100.00 (1 leg) | unlabelled EOA |
| **TOTAL** | **$32,505.61** | **about $2,322/day out of his own pocket** |

**So roughly $23,485 of the $28,771 "TVL" is the operator's own USDG, put in over the last two weeks.** External adoption is close to nil, and the earlier "$5,000 from @toady_hawk is the whole external adoption" read still broadly holds.

## Second contract stack, deployed 5 Sep. They ARE shipping, and it is verifiable on chain, not just claimed.
AllowlistedFactory `0xa9ab194FB74dFD9991047839aE23A576c8403d95` · OperatorControllerV2 `0x705F4f94344feD3f7BA98300F8bC04fdAE570340` · VaultAllowlist `0x3B1B0f1812a9a09664D82238E94c604D30Bb8134` · Registry `0xdF1a23B1A7507Cc3B270DfA78FDD9ddA7bC36325` · PoolZapper `0x8934A298d3E59a632994ac254Dee19112127f2B7` · range runner `0x372f6aedaba9989186e52B1E9d2e7Ea5572aa6cF` · V4Quoter `0x8Dc178eFB8111BB0973Dd9d722ebeFF267c98F94`.
**Governance also shipped:** VoteToken vTWO `0x5c02401e945d86FB7109EC77F358498D6DA05950` (wrap TWO 1:1, no owner) and PoolVoting `0x6FDcc4A912b0f7d73251b7a719Aa2943734fD6E7`. GitHub could not be checked (the container's GitHub API is repo-scoped), but a full second stack plus governance on chain is stronger evidence of shipping than a commit graph.

## Live market, 18 Sep 2026 (it kept falling DURING this audit)
**$0.0001993 · mcap = FDV $199,315 · liq $56,788.70 · vol24h $20,129 · 68 buys / 45 sells · -10.12% in 24h.**
Earlier the same morning it was $0.0002540 / $254,053 / $65,009. **Down another 21.5% in a few hours.** Do not quote a stale mcap on this name; re-read before every statement.
Third-party bulls still posting: @MedbdyLLC (5,165f, 12 Sep, "Still holding. Still the best yield setup we've seen on RH") and @toady_hawk (19,728f). Also a "vote to list $TWO on the Robinhood Top 10" campaign from small accounts.

## REVISED VERDICT, 18 Sep 2026: small speculative yes, and the reason changed
The earlier NO rested on two things that turned out to be false: that staking had collapsed, and that the founder was still distributing. **Neither survived the audit.** What is actually true:
- The founder is **net +10.7M TWO** over 14 days and is spending **$2,322/day of his own USDG** to keep the product alive at a $199,315 market cap. He is putting money in, not taking it out. At this cap that alignment is the whole bull case.
- The 13.07% free handout (see [[twofold-two-2]]) is **fully distributed and gone**. Those wallets are empty. That overhang cannot repeat.
- LP permanently locked, token renounced, second full contract stack plus governance shipped on chain since 5 Sep.
Against that: **revenue is $11.17/day**, TVL is mostly the founder's own money, and the documented harvest-to-stakers path has paid $0.17 ever. This is a bet on an operator, not on a business.
**Position: lottery-ticket size only.** The thing that kills it is the subsidy stopping.
**Falsifiers, all one API call or one RPC call:**
1. Operator USDG outflow drops toward zero for a week → he has stopped funding it, exit.
2. `totalStaked()` on `0x9CF18bB1dD9AfBF75B579Cc0C473B2975c16E9e3` falling while the 61.7M gap drains → the 7-day queue is emptying into the market.
3. Bull confirmation: own-pool fees at `twofold.fi/api/pools` above **$100/day** (now $11.17), or `users7d` at `/api/stats` back above 200 (now 50), or OperatorControllerV2 harvest legs going from $0.17 to something real.

# SECOND TRIPLE-CHECK, 18 Sep 2026. Ben asked again, plus a full X claim audit. Three more of my own claims broke.

## CORRECTION 5, and it is the big one: "the founder burns $2,322/day of his own money" was WRONG. I measured GROSS.
That was the single load-bearing fact under my "small speculative yes". It does not survive.
Operator EOA `0xC8720447712e6C4c851B3884b4Ec93F9cE8aD5fD`, USDG, 14 days:
- **OUT $32,505.64 in 1,077 legs · IN $34,988.71 in 67 legs · NET +$2,483.07 IN HIS FAVOUR.**
- The money comes back from his own pools: $11,943.44 from hook 1 `0x127b3f3b7769f659c5edbff8b4005443f19faac0`, $7,991.59 from hook 2 `0xd1bcbcca41f3bdb6b4812652959c6df725ea2ac0`, $8,690.04 from EOA `0xf70da97812cb96acdf810712aa562db8dfa3dbef`, $5,121.88 from `0xbdbae060cbab0e9cfe802a7513dd5ecb36cda6c3`, $1,240.66 from router `0x8f10b468b06c6fd214b65f87778827f7d113f996`.
- WETH over the same window: out 8.0385, in 7.9565, **net -0.0819 WETH**. Immaterial.
**He is running roughly FLAT, not subsidising.** He seeds pools, pulls capital back, and recycles some into staking rewards. **This is the standing "measure cumulative inflow, not current balance, and check the outbound legs" rule, broken again by me. Gross outflow is never a burn rate.**

## CORRECTION 6: "the founder handout wallets are empty" was WRONG. I forgot Vault 2.
I checked `balanceOf` plus `staked()` on Vault 1 only. Two of the seeded wallets are staked in **Vault 2**:
`0x37c35cd23f5d91ac5158992aa29144a804e2f54d` liquid 1 TWO but **18,936,820 staked**; `0xfa7617537cff17c28ed13706509541220631f82a` liquid 288,666 but **15,007,222 staked**.
**Founder cluster (operator + 7 seeded wallets) still holds 61,637,892 TWO = 6.16% of supply (about $12,300).** Not gone. This is the third time on this token that a zero `balanceOf` was mistaken for an exit. **Rule, now absolute: enumerate EVERY staking contract a project has before calling any wallet empty.**
Neither vault holds operator stake: `staked(operator)` = 0 on both. Of Vault 2's 116,538,217, **33,944,042 (29.1%) is founder-seeded wallets and 82,594,175 (8.26% of supply) is independent**.

## CORRECTION 7, in the operator's FAVOUR: he has never sold. Proven, whole life.
I had said he "sold 7,495,572 TWO into the pool" and then hedged. The hedge was right and the claim was wrong. Pulled **every** Transfer of TWO from the operator EOA to the PoolManager since genesis, then read each transaction receipt for `Swap` (topic `0x40e9cecb…`) versus `ModifyLiquidity` (topic `0xf208f491…`) and for what came back to him in the same tx:
- **11 legs, 853,136,018 TWO total. 8 are LP ADDS** (ModifyLiquidity, nothing back), including the 841,500,000 genesis add on 28 Aug 23:04 BKK, tx `0xf3f5041f64232ca26587fee590c2e622e808faced80fc109e470a97a9194de3a`.
- **3 are sells, all on 30 Aug 10:24 BKK, 1,000 TWO each, returning USDG 2.8773 + 2.8740 + 2.8707 = $8.62 total.** Test transactions.
**@twofoldfi's 9 Sep claim "The deployer has never sold TWO, into our pools or anywhere else" is TRUE in substance** (false by $8.62). This is the strongest single fact in the token's favour and it is now proven rather than assumed.
**Method note for onchain-forensics: on v4, resolve a PoolManager transfer by reading the receipt for Swap vs ModifyLiquidity and for the quote asset returning to the sender in the same tx. Never label it from the Transfer alone.**

## X CLAIM AUDIT, @twofoldfi (2,028 followers), full post and reply history
| claim | date, views | verdict |
|---|---|---|
| "The deployer has never sold TWO, into our pools or anywhere else" | 9 Sep, 122 | **TRUE** (see above, $8.62 of test sells) |
| "DualPool is on the Uniswap hook routing allowlist" | 9 Sep, **17,100** | **UNRESOLVED.** Uniswap's hook-registry repo 404s; their quote gateway 500s on chain 4663. Could not verify either way. |
| "DualPool adapter is live in the @KyberNetwork aggregator" | 9 Sep, 3,762 | **PARTLY TRUE, effectively FALSE.** Kyber does serve Robinhood Chain (`aggregator-api.kyberswap.com/robinhood/...`, chain 4663 present in its config) and does price TWO. But I requested routes at $1, $100 and $1,000 of USDG→TWO: **every route goes through non-Twofold pools** (`0xb8d83170ad5e…`, the no-hook TWO/WETH pool; `0x4bd4a4d34618c082ba01ff53b971c7d97a05cd9f865ee29b7b54b2fe7c4b7252`; `0xde26bb6272471c6570821d3027afd7f26b388454e3d1296abbc0051b9bdb6b2e`), **none of which is in Twofold's own 64-pool list.** Exchange types returned are only `uniswap-v4`, `uniswap-v4-fee`, `pmm-18`; neither hook address appears anywhere in the response. Real order flow reaching Twofold from Kyber right now: none detected. |
| **"The TWO/USDG pool counts double… That pool is where the aggregators route TWO trades, so the swap fees are real on top"** | 10 Sep, 589 | **FALSE, and this is the cleanest falsification of the audit.** Twofold's three own TWO/USDG pools hold **$79, $1 and $1** and did **zero swaps in 24h**. The double-weighted pool in their own `/api/rewards` `poolWeights` is `0x9699c1059e66385e25692a8b8e7297e0cc8929cdcc9cc1117fa242281b13f85b`, **the $1 one.** |
| "We pay rewards from trading fees the protocol's own positions earn. Nothing is minted. No TWO gets sold to fund it." | 10 Sep, 436 | **FALSE on the first clause, true on the rest.** The staking vault's USDG came 100% from the operator EOA ($6,355.00 in 952 legs). Protocol swap fees are $11.17/day against $316.59/day of payouts, so fees cannot be the source. Nothing is minted and no TWO is sold, both true. |
| "USDG vault live and paying, over 290% APY" | 17 Sep ×2 | **TRUE and if anything understated.** $316.59/day on 116,538,217 TWO staked (about $23,482) is roughly 490% at today's price. |
| "37 pools now live, up from 29" / "Completely new contract stack" | 6 Sep, 6,384 | **TRUE.** 64 pool ids listed today across two stacks; the second stack is on chain (see part 3 above). |
| "Payouts every day… a Paid event on chain from the RewardsPayer contract" | 10 Sep, 662 | **TRUE.** 246 payout legs to 68 distinct wallets in 14 days. |
| "Tokenized US stocks just got a regulatory path onchain… open to you" | 17 Sep, 1,166 | Rides [[sec-innovation-exemption]]. Their pools are open; the regulatory framing is theirs, not verified here. |

## OPERATOR QUALITY: they block people, including their own bulls
Independently reported by four accounts on 12–13 Sep: **@MedbdyLLC (5,163f), who posted "Still holding. Still the best yield setup we've seen on RH" (3,603 views), was blocked.** @toady_hawk (19,729f) asked publicly "did you block this bull by mistake perhaps?". @Adebayo2056: "they blocked me too I quoted saying bad price action but good tek and I was blocked". @KrystianDeFi and @PPFUN8 corroborate.
Other outside voices: @TNr1ck (4,459f) positive, @0xlife_ai (1,148f) positive on design, @IvanBullish (14,988f) neutral chain-level mention, **@0xYelf (6,708f) "shitcoin"**, **@ChillTRD (33,689f) "L"**, **@bukka_ram16849 "scam"**. @baseyieldguy asked the right question and got no answer: *"$500k MC with a merged Uniswap hook is either a market asleep or a red flag i'm not seeing. what's the TVL look like?"* The answer is $28,771.
**Tier-one coverage still zero.** blknoiz06's 5 hits are the English word "twofold" in unrelated posts; theunipcs, pingucharts, CrashiusClay69, 0xkioto all return 0.

## FINAL VERDICT after two rounds of triple-checking: NO. Reverting the pivot.
The "small speculative yes" rested on the founder spending his own money. **He is net +$2,483 over 14 days. That fact is gone, so the pivot goes with it.**
What survives in its favour, all verified: **he has never sold** (proven, whole life, $8.62 of test trades), genesis LP of 841,500,000 TWO is permanently locked in `0x485690D46e344127aa2EB45D8d9BAcbb5856D58b` with liquidity byte-identical since 31 Aug, the token is renounced with no mint/pause/blacklist, a full second contract stack plus vTWO governance shipped on chain, and 82,594,175 TWO of independent stake sits behind a 7-day cooldown.
What kills it: **revenue $11.17/day**, TVL **$28,771**, the flagship TWO/USDG pool holds **$1**, the aggregator-routing claim that drew 17k and 3.7k views does not show up in any route I could pull, and the team blocks people who ask about price while holding the bag. Price $0.0002015, mcap **$201,549**, still down about 9.4% in 24h and it fell ~21% during the first audit.
**This is a well-built machine with no customers whose marketing overstates its distribution. That combination does not re-rate on a founder being honest about not selling.**
**One trigger to revisit, one call: a Kyber route at `aggregator-api.kyberswap.com/robinhood/api/v1/routes` that actually contains a Twofold pool id, together with own-pool fees above $100/day at `twofold.fi/api/pools`.** Until a real route lands there, the product has no distribution and the market cap is not "low for what it does", it is priced for what it does.

# THIRD TRIPLE-CHECK, 18 Sep 2026. The three numbers the NO verdict rests on, each verified INDEPENDENTLY of the project's own API. All three hold.
I attacked my own three load-bearing numbers, because each had until now come from `twofold.fi/api/pools` rather than from the chain.

## Iteration 1 — REAL TVL, via ERC-4626, the check memory prescribes and I had never run
DualPools are JIT: the USDG rests in a Morpho vault, not in the pool, so `reserves` could in principle have hidden most of the capital. Ran `convertToAssets(balanceOf(hook))` on all four allowlisted Steakhouse vaults for both hooks:
- hook 1 `0x127B3f3b7769f659C5eDBfF8b4005443f19FAAc0`: 1,653,112,568,198,217,574,348 shares of Steakhouse USDG `0xBeEff033F34C046626B8D0A041844C5d1A5409dd` = **$1,665.17**, loose USDG $0.00
- hook 2 `0xd1BcbCCa41f3bdb6b4812652959c6dF725ea2Ac0`: 12,272,363,444,961,482,475,219 shares = **$12,361.84**, loose USDG $0.00
- Zero shares in the Ethena, Turbo and Grove vaults.
- **Total USDG-side capital $14,027.01.** Two-sided that is about $28,000, against the $28,771 I had from their API. **No hidden TVL. My figure holds.**

## Iteration 2 — REAL REVENUE, from on-chain v4 Swap events, not their API
Swept `Swap` (topic0 `0x40e9cecb9f5f1f1c5b9c97dec2917b7ee92e57ba5563708daca94dd84ad7112f`) on the PoolManager filtered to all 64 Twofold pool ids, 24h, and read the **`fee` field out of each event** (word 6) rather than trusting a configured tier.
- **1,316 swaps on chain** (their API claimed 1,472). Fee tiers actually charged: 100 ×1032, 500 ×130, 3000 ×101, 5000 ×46, 2000 ×7.
- **Total volume $11,207.52, total swap fees $9.43 in 24 hours.** I had reported $11.17/day from their API. **The real number is 16% LOWER. My figure held and was, if anything, generous to them.**
- **Composition matters: 1,039 of the 1,316 swaps are the single WETH/USDG 0.01% pool, earning $0.80.** That is arb/routing traffic, not product usage. The actual equity DualPools do about 270 swaps a day between them. MEME/USDG ($5.23) and PONS/USDG ($1.90) are 76% of all fee income.

## Iteration 3 — the Kyber test redone properly. My earlier test was methodologically weak.
I had only tested USDG→TWO. That was the wrong pair: Twofold's own TWO pools hold $79, $1 and $1, so of course nothing routes there. Re-ran on the pairs where they actually have depth, $100 each:
| pair | route Kyber picked | Twofold pool in route? |
|---|---|---|
| USDG→NVDA | uniswap-v4-fee, then uniswapv3 `0x62ab521f71431f78ac374cdbadc6cda3c8916b6c` | **no** |
| USDG→TSLA | uniswap-v4-fee, then **alandale-v4** `0x352771b68da27e3bece96eb2c8532d7faed1de1b` | **no** |
| USDG→AAPL | uniswap-v4-fee, then uniswapv3 `0x8bb3514e2204e1cdf3ac149efee7ff04d91b719f` | **no** |
| USDG→SPY | uniswap-v4-fee, uniswap-v4-passthru, uniswap-v4 | **no** |
| USDG→INTC | uniswap-v4-fee, **ramses-v3**, uniswapv3 | **no** |
**Eight independent route requests now (TWO at $1/$100/$1,000 plus these five), zero Twofold pools in any of them.** Decisive detail: Kyber's response carries named adapters for **alandale-v4, ramses-v3, uniswapv3, uniswap-v4, uniswap-v4-fee, uniswap-v4-passthru and pmm-18**. It knows Robinhood Chain DEXs by name, including a direct competitor. **There is no twofold or dualpool exchange type in any response.** The 9 Sep "DualPool adapter is live in the KyberNetwork aggregator" post (3,762 views) produces no observable routed flow, and the 10 Sep "that pool is where the aggregators route TWO trades" (589 views) is flatly false.

## Operator's latest activity, 18 Sep
- Nonce 4,299. **Last LP add was 8 Sep** (see the 11-leg table above). He has not seeded a pool in 10 days.
- **USDG out of his wallet in the last 3 days: ONE leg, $1,600.00, on 17 Sep 04:53 BKK, to TwoStakingUSDG `0x9CF18bB1dD9AfBF75B579Cc0C473B2975c16E9e3`.** He has switched from 952 dribble legs to lump-sum top-ups. Eight days of rewards at the $200/day ceiling. Still funding, less often.

## Price through this session, all Bangkok time 18 Sep
$0.0002540 / mcap $254,053 → $0.0001993 / $199,315 → $0.0002015 / $201,549 → **$0.0001949 / mcap $194,964, liq $56,119, −12.67% in 24h, 67 buys / 47 sells.**
**Down 23% over the few hours of this audit alone.** Never quote a stale mcap on this name.

## VERDICT UNCHANGED AND NOW FULLY INDEPENDENT: NO.
Every number in the bear case has been re-derived from the chain rather than from the project:
- revenue **$9.43/day** (on-chain Swap events)
- capital **$14,027 USDG-side** (ERC-4626 convertToAssets)
- distribution **zero routed flow** (8 Kyber route tests)
- and 79% of the swap count is one arb pool earning $0.80.
The founder is genuinely clean on the one thing that usually kills these (he has never sold, proven whole-life, $8.62 of test trades) and the LP is genuinely locked forever. **But clean is not a business.** The thing Ben liked, "an extremely low market cap for what the project is doing", does not survive measurement: what it is doing is $9.43 a day, and its own headline distribution claim is not observable in any aggregator route.
**Revisit trigger, unchanged and now precisely specified: one Kyber route at `aggregator-api.kyberswap.com/robinhood/api/v1/routes` whose `route[][].pool` matches a poolId in `twofold.fi/api/pools`, plus on-chain 24h Swap fees above $100/day excluding the WETH/USDG 0.01% pool.**

# 18 Sep 2026 — STANDING METHOD CORRECTION. Ben challenged the whole framework, and he is right.
He said: "I just want to make sure you're not making the same mistake you made with HOOKR and DELTA."

## The HOOKR miss, owned
**15 Sep I wrote HARD NO on HOOKR at $0.007944 / $7.92M cap. On 18 Sep it is $0.01500 / $14,933,887, +89.54% in 24h, +18.25% in 6h, vol24h $1,658,402, 1,970 buys / 2,005 sells.** Roughly +88% from the call.
**I re-verified every fact behind that HARD NO today and every one still holds:** HookrFlywheelBurner `0x8cee20fa000af3266ac2cd2cbeefbcd19d98fd89` holds **0.036107 ETH, the identical number to 15 Sep** (never fired); ZAPS burner `0xb5f7d9d4269c897df70df26f7ba48c0d933be8db` still 0 ETH; hook ERC-6909 ETH claim **0.152306 vs 0.152106 on 15 Sep**, i.e. +0.0002 ETH in three days; overhang EOA `0x3386b004ec4a2fc7cbdef7f933ef98e83a9c2fcd` still exactly 1,915,891; deployer 23,302,928 → 23,095,684; dEaD 4,427,127 → 4,704,807 (inventory burns again, not market buys).
**So the facts were right and the trade call was wrong. Nothing about the cash flows changed. What changed is that @Uniswap publicly engaged with them.** Traders name it explicitly: @89trading (55,868f) "$hookr đã vượt qua $frong ... nhờ comment của Uniswap", @Intel_degen "Maybe $hookr? Uniswap already started talking about them", @lbexplorer (10,684f, 3,036 views) "Study J-curve $HOOKR and $DELTA coded. Both up 2x since", @im_goomba (64,983f) "on the comeback trail after the fud". @NodarJ's 17 Sep post did 17,651 views / 118 likes.

## Ben's DELTA recollection is NOT right, and the record should be quoted back
He said "You said that Delta isn't making any money." **The file says the opposite, and the one time I did write it, he caught it and it was fixed the same day.**
- I wrote "no revenue line on DefiLlama at all" on a 6 Sep bottoming screen. **That was the error, it is already logged in [[delta-liquidity-2]] with a standing gate: never write "no revenue" until the treasury measurement has been RUN (`/tmp/bt/deltarev.py`, 4 minutes); otherwise the honest output is "not measured this pass".**
- What the file actually concluded, repeatedly and with measurements: revenue **$244,000/yr (31 Aug) → $576,194/yr (4 Sep) → $1,131,604/yr → $1,315,337/yr (6 Sep)**, cap/revenue **120x → 36.5x → 9.7x → 8.3x**. Verbatim: *"The revenue is real and the order of magnitude is settled."*
- My actual claim was narrower and was true at the time: the take was ON at 1% (`feeBps()` = 100, verified three ways) but **no route existed to holders** (fee recipient `0x266311a44e70363b3d168bcb38cf38cbf2ccc51f` had no distribution code).
- **I set the trigger and the trigger worked.** 31 Aug: *"TRIGGER TO FLIP TO YES: feeBps() above 100, or any DELTA-denominated buyback appearing on chain."* On 7 Sep 22:14 UTC both landed (new manager `0xbCb96b15dC2246D242c879316c86e25e846ad5FB` feeBps 750, first buyback-and-burn) and **the verdict flipped HOLD → BUY at $0.01354 / $13,508,782**, then STRONG BUY from the bot on 8 Sep. **DELTA is $0.01457 / $14,506,431 today.** That one was called correctly.

## DELTA burn watch RESOLVED, and it is good news
The 17 Sep entry flagged that the 3-day burn clock (7, 10, 13 Sep) had missed a slot and that **20 Sep was the date that settled it**. Checked live today: **dEaD balance 5,004,654 DELTA, up 684,363 since 13 Sep. Burn #4 landed. The program is alive.** Treasury DELTA balance 0.00 (it buys and burns rather than accumulating, as expected). `feeBps()` = **750 on BOTH** managers `0xbCb96b15dC2246D242c879316c86e25e846ad5FB` and `0x83551a233f6d7b3cb4defbf55abfd887c6811ce4`. **The 13 Sep falsifier has NOT fired. Close the 17 Sep worry.**

## THE ACTUAL LESSON, and it is a standing correction to how I answer Ben
**On Robinhood Chain, protocol revenue has near-zero predictive power over a 1-to-4-week move. Attention does.** This is already written in my own notes and I keep overriding it: the "pays the holder automatically" bucket ran **+1,925% on a 30-day median** while "infrastructure the holder must operate" ran **+29%** — 66x apart, on category, not on cash flow.
**I have been answering a trading question with a valuation framework.** Revenue and cap/revenue tell you what a thing is worth if it is ever priced on fundamentals. On this chain, in this regime, they do not tell you what it does over the next month. HOOKR is the proof: identical cash flows on 15 and 18 September, +89%.
**NEW RULE, applies to every screen from here: lead the verdict with the ATTENTION vector, not the revenue line. Revenue belongs in the valuation frames; it must never again be the primary reason for a yes or a no on a 1-to-4-week horizon.** The attention vector is: does any account above ~50k followers hold or post it, is a protocol with real distribution (Uniswap, Kyber, a launchpad) engaging publicly, and is the holder count rising.

## SO, IS THE TWO "NO" THE SAME MISTAKE? Partly yes. The verdict survives but the REASON has to change.
**The old reason was wrong-shaped:** "$9.43/day of revenue" was never going to predict TWO's next month, exactly as it failed on HOOKR.
**The reason that actually holds up, and it is the one difference that matters:**
| | HOOKR | TWO |
|---|---|---|
| revenue | ~$0 reaching the token | $9.43/day |
| tier-one coverage | @Bluntz_Capital 338,562f, @ripchillpill 99,604f, @_0xghost_ 44,479f, @DineroDom0, **and @Uniswap itself** | **zero, ever, re-verified three times** |
| followers | @Hookrfun 2,517 and a live community | @twofoldfi 2,028, and **they block their own bulls** (@MedbdyLLC, 5,163f, publicly holding, blocked) |
| holders | 6,792 and rising through a drawdown | falling |
**TWO has no attention vector at all, and it removes the ones it gets.** That is the reason to pass, and it is a far better reason than the fee number.

## Also correcting my own overreach from the pass above
I called the Kyber claim "PARTLY TRUE, effectively FALSE". **Too strong, and I am walking it back.** Twofold's pools hold about $12,000. **A router would correctly never select a $12k pool for a $100 trade even if the adapter is perfectly live.** Non-selection is fully explained by depth, so my eight route tests do NOT prove the adapter is absent. Honest version: **I cannot confirm or deny the adapter exists; what I can say is that no routed flow arrives today.**
**The claim that IS flatly false stands: "That pool is where the aggregators route TWO trades" — that pool holds $1 and did zero swaps in 24h.**

# 18 Sep 2026 — "is TWO one caller away from pumping?" THE DECISIVE FINDING, and it was hiding in plain sight.
Ben's thesis: everything is there, they just lack attention. **Half right mechanically, wrong structurally, and the structural half is now proven.**

## The mechanical half is RIGHT and the numbers are small
Live: **$0.0001909, mcap $190,934, aggregate liquidity $57,359 across 4 pairs, vol24h $20,964, 67 buys / 49 sells, −14.47% 24h, −20.53% 6h, −5.06% 1h.** Still falling and accelerating.
Net buy required, constant product, an UPPER bound because v4 concentration makes it cheaper:
| target | net buy | resulting mcap |
|---|---|---|
| 2x | **$11,879** | $381,868 |
| 3x | $20,995 | $572,802 |
| 5x | $35,450 | $954,670 |
| **10x** | **$62,013** | $1,909,340 |
| 20x | $99,578 | $3,818,680 |
| 50x | $174,114 | $9,546,700 |
**$11,879 doubles it. One mid-tier caller's audience could do that in an hour.** Ben's instinct about the powder keg is correct, and the 7-day unstake cooldown on Vault 2 (116,538,217 TWO) means a large part of the float cannot react quickly, which makes a squeeze sharper.

## THE STRUCTURAL HALF IS WRONG, AND THIS IS THE FINDING: **DualPool is UNISWAP'S OWN HOOK. Twofold did not invent it.**
Pulled @Uniswap's own timeline. **22 July 2026, a full thread, verbatim:**
- *"DualPool hook was designed and built in collaboration with @sparkfinance. Here's how it works: → Deployer parks assets in an ERC-4626 vault → The hook pulls in liquidity…"*
- *"DualPool hook is audited, open source, and ready to implement. **Any team can now deploy their own DualPool hook**, allowing them to earn on both active and idle assets"*
- *"Tune in for a livestream with @haydenzadams and @hexonaut, where they'll discuss the DualPool hook, **Spark's $150M liquidity migration**, and the stablecoin FX layer"*
- Plus *"DualPool hook developer docs ↓"*.
@saintniko (25,466f, Uniswap-affiliated), 25 June: *"spark just migrated $150M to uniswap. up next, dualpool hook (which will allow them to earn on both active and idle capital)"*.
**Twofold's own docs already said this and I underweighted it: "DualPoolHook, Uniswap's audited v4 hook, deployed unmodified."**

### What that means, and it is the answer to Ben's question
1. **The attention already happened.** The DualPool mechanism got Uniswap-scale attention in July: an official thread, developer docs, a Hayden Adams livestream, and a $150M Spark migration. **It flowed to Uniswap and Spark. Twofold got exactly one mention: @saintniko, 31 Aug, "ahh so cool! they're using the dual pool hook!!!!"** — praise for the HOOK, addressed to @toady_hawk, not a call on the token. Memory already recorded that nuance on 2 Sep and it is now fully explained.
2. **Twofold has no moat and Uniswap says so in writing: "any team can now deploy their own DualPool hook."** Anybody can be Twofold tomorrow, with Uniswap's audited code, for free.
3. **This is the exact structural difference from HOOKR, and it is why the HOOKR analogy fails.** HOOKR *is* the hook marketplace, so when Uniswap says "hooks are eating the world," HOOKR is the only tradeable proxy for that sentence and nobody else occupies the slot. **Twofold is one of N deployers of somebody else's hook. When DualPool attention arrives it routes to Uniswap and Spark, and Twofold is a customer of the narrative rather than its owner.**
4. Big-account history, all time, on `(twofoldfi OR dualpool)`: only **2 of 20 posts from accounts ≥15k followers** — @BiteyeCN (80,300f, a bare link, 6,020 views) and @PinkBrains_io (15,138f, an ecosystem map listing many projects, 9,495 views). Neither is a call.
5. **Method note, the blknoiz06 trap again and I nearly repeated it:** the raw counts were Uniswap 11, saintniko 3, haydenzadams 1. **Hayden's was the English word "twofold" and one of Uniswap's was a bare link.** Always print the matched text and mark real hits before reporting a count.

## Honest steelman for Ben's side, because it is not nothing
A caller does not need to own a narrative to run a coin. **"Uniswap's own audited hook, live, at a $190k market cap"** is a genuinely sellable line, it is TRUE, and $11,879 doubles the thing. Add a locked LP, a renounced token and a founder proven never to have sold, and the story writes itself. **If a 50k-follower account ever does pick it up, this moves violently.** That path is real; it is just not the same path HOOKR took, and nothing in the evidence says anyone is preparing to walk it.

## VERDICT HOLDS: no. Reason restated in its final and best form.
Not "it earns $9.43/day". Not "the founder is dumping" (he never sold). **It is that Twofold rents its story from Uniswap, and the rent has already been collected by Uniswap and Spark.** The one asset a $190k micro cap needs is a story nobody else can tell, and Uniswap open-sourced this one in July.
**The trigger to revisit is unchanged in spirit and sharper in form: a named account above ~50k followers posting the CONTRACT ADDRESS with a thesis, not the mechanism.** Mechanism praise already exists and moved nothing.