Monday edition — reviews Sep 15–21, sets the agenda for the week of Sep 21. Robinhood Chain did 142 ETH ≈ $389K and 69 launches — down −91% / −69% from last week's corrected 1,531 ETH / 223. That collapse landed eight days before RHC's free-gas incentive ends Sep 29: the volume is leaving before the subsidy does. Solana held: prior full week $1.43M (−24% off the stock-pairs spike), then Monday Sep 21 alone printed $619K — the strongest day in a fortnight. Eng spent Monday on one thing that matters: fixing "user can't claim" — versioned creator-claim transactions + v0 lookup-table support shipped across backend and web. And a repeat of last quarter's lesson: our RH scanner was blind again (stale token cache → a false "0 launches"); corrected here. This is a "prove the engine still runs" week.
Operating frame: Bags runs as a KPI-focused firm aimed at capturing trading volume through mobile and terminal products. Stats come prefilled (this brief is the prefill); the meeting discusses why. This week the "why" is stark: Robinhood volume dropped 91% with the free-gas cliff a week away, while Solana printed its best Monday in two weeks. One chain is fading on schedule; the other just reminded us it can still spike.
Key stats: RH 142 ETH ≈ $389K (−91% WoW), 69 launches (−69%), fees ≈ $7.8K — a full-scale collapse, 8 days before free gas ends. Solana prior week $1.43M (−24% off the stock-pairs peak), 276 launches (−1%), 6.3K traders; then Mon Sep 21 = $619K single-day. Eng (Mon only): ~11 human commits, creator-claim fix shipped (SUPPORT-35). Public revenue: DefiLlama 7d $26.6K (Solana-only; RH is additive on top). Headline: Robinhood was 88% of take last month — this week it's collapsing, and Solana is the story again.
Solana (prefilled): prior full week $1.43M volume (−24% WoW off the stock-pairs spike week) · 276 launches (−1%) · 220 creators (0%) · 6.3K traders (−14%) · 24.6K trades (−31%). Then Monday Sep 21 = $619K, 21 launches, 1,887 traders — a genuine pop, the best day since the stock-pairs announcement. Stock pairs cooled hard: 5 of 276 launches (1.8%) vs 30/156 (19%) last week — the experiment lost momentum the moment the announcement bump faded. Creator claims (prior wk) $15.5K / 175 claimers.
Robinhood (corrected scan): 142.38 ETH ≈ $389K (−91% WoW), 69 launches (~10/day, −69%), fees ≈ $7.8K, protocol half ≈ $3.9K. Curve 104 ETH / pool 38 ETH — pool volume all but died (it was 1,090 ETH last week). Two curve spikes (Sep 15: 47 ETH, Sep 17: 42 ETH) then a slide to 6 → 4 ETH by the weekend. Vault swept to empty ($0.27). All-time launches 4,673.
The Monday story was claims, not features. Backend #1302 "fit creator claims with v0 lookup tables" + tc #532/#533 "support versioned creator claim transactions" shipped to fix SUPPORT-35 "user can't claim" — a revenue-blocking bug for creators, closed same-day. In flight: a big terminal/web queue — tc #535 Robinhood v3 launches with stock-pair + trade-fee (+7,842), #531 BagsPay shell v3 RH + Axiom (+19,574), #534 launch-form redesign, #530 /transparency page, #529 deployer revenue claims. Mobile: Finn's bagsapp stack is queued — #675 fomo feed + grok chat (+2,752), #671 discovery refresh (+2,309), #661 Top Traders (+736), and the 96-day #649 Cashback/Hyperliquid perps monster (+25,618). Top Traders #661: week eight.
Data / indexer queue is heavy: alaa's backend #1280 "native quotes through fee-share utilities" (+55,255) + datapi #145 "exact quote fee analytics + valuation" (+9,708) + indexer #142 "native quote fee-sharing facts" (+43,273) — a large native-quote fee-accounting push in review. #153 MetaDAO/Futarchy + #152 Candle launchpad extend coverage; #148 recovers wallet positions from unpriced routed legs. Auth: backend #1281 "authentication re-audit + engineering handoff" (+115,392 docs) + #1259 wallet verify + wallet-verify repo #2 — a real auth hardening effort spinning up. bags.automations shipped fee-bot ECR/OIDC + a CI stale-image guard (#134). AMM: status in-meeting.
Three items that can't wait: (1) The RHC cliff — free gas ends Sep 29 and volume already fell 91% ahead of it. Decide the posture now: do we do anything to retain RH activity post-subsidy, or accept RH becomes a thin tail and re-center on Solana? (2) The Solana take-rate anomaly (again) — the platform-take table shows $16.7K accrued in 2 days this week vs $4.8K for all of last week; that's a ~2.6% implied Monday take rate, almost certainly batched/lagged accruals landing, not real. Nobody quotes a Solana take rate until eng reconciles the fee path. (3) The scanner blind spot recurred — same stale-cache bug as last quarter produced a false "0 launches / 39 ETH." Fixed by refreshing the cache; the scanner needs a real fix so this doesn't burn a third deck. Close with named owners.
This is a Monday-night brief — the GitHub window is one day, so the commit counts are small by construction and the full week will be larger. What shipped Monday is telling: the whole team's energy went to unblocking creator fee claims (versioned transactions + v0 lookup tables) — the exact thing SUPPORT-35 flagged. Everything else was infra: fee-bot deploy plumbing and a CI guard against unsafe image promotions.
Green = human GitHub commits (Mon Sep 21 only; deploy[bot], renovate, cursor excluded). Linear: 1 issue closed — SUPPORT-35 "User can't claim" (ramyo) — which maps 1:1 to the day's commits. Monday-only window: the who-shipped picture fills in over the week; the signal here is focus, not volume — the whole team pointed at one revenue-blocking bug and closed it.
| Theme | What landed | Owner | Repo |
|---|---|---|---|
| 🩹 Creators can claim again | Backend #1302 "fit creator claims with v0 lookup tables" + tc #533 claim fix + tc #532 "support versioned creator claim transactions." Fixes SUPPORT-35 "user can't claim" — versioned (v0) transactions with address-lookup-tables weren't being handled on the claim path, so a class of creators couldn't withdraw fees. Diagnosed and shipped across backend + web the same day. | ramyodev | bags.backend · bags-twitter-claim |
| ⚙️ Fee-bot + CI plumbing | Fee bots ECR + OIDC (deploy identity for the fee-collection bots), staging backend image bumps (1134a6c, 7f99e73), production backend → 1.49.1, and #134 "restore stale-image warning in promote-image, fail on unsafe promotions" — a CI guardrail so a stale image can't be promoted to prod silently. | ramyodev | bags.automations |
| 🗂 Indexer backfill wiring | #116 "mongo wiring and control-plane artifacts" merged — backfill control-plane plumbing for the indexer. | alaabags | solana-indexer-v2 |
Read: a focused Monday. When a creator-payout bug surfaces, the correct response is to drop everything and fix it — that happened. The rest of the week's shipping (the RH v3 / BagsPay / native-quote-fees queue in the next slide) is where the volume of work actually sits.
The open queue is unusually heavy on large, structural PRs. Three clusters dominate: (1) native-quote fee-sharing across indexer + datapi + backend (100K+ lines combined) — the accounting rework that should also fix the Solana take-rate anomaly; (2) Robinhood v3 + BagsPay on web (+27K lines across two PRs); (3) an auth re-audit. Meanwhile Finn's mobile stack (#675/#671/#661/#649) keeps aging.
| PR | Scope | Owner | Status | |
|---|---|---|---|---|
| HOT | indexer #142 · backend #1280 · datapi #145 — native-quote fee-sharing | +43,273 / +55,255 / +9,708 — index native-quote fee-sharing facts, push them through the fee-share utilities, and expose exact quote-fee analytics + valuation. This is the accounting rework that likely resolves the take-rate anomaly. Biggest thing in the org right now. | alaabags | in review |
| HOT | twitter-claim #535 · RH v3 launches (stock-pair + trade-fee) | +7,842 — Robinhood v3 launch flow with stock-pair support and the trade-fee pipe. This is the RH fee-capture work; it matters more if we decide to keep investing in RH past the Sep 29 cliff. | finnbags | open |
| HOT | twitter-claim #531 · BagsPay shell v3 RH + Axiom | +19,574 — the BagsPay surface + v3 Robinhood launches + Axiom trade terminal. Largest web PR open; needs to be broken up or landed deliberately. | finnbags | open |
| HOT | backend #1281 auth re-audit · #1259 wallet verify · wallet-verify #2 | +115,392 (docs-heavy) auth re-audit + engineering handoff, plus wallet-verify reborn (backend #1259 + nils' stateful wallet-verify repo). A coordinated auth/identity hardening push. | ramyo · nilsbags | Sep |
| WEB | tc #534 launch-form redesign · #530 /transparency page · #529 deployer revenue claims · #525 swaps→trade v2 · #513 apps page as app store | The web-terminal queue: a launch-form redesign, a public /transparency page, deployer revenue claims on profiles, the swaps→trade-v2 migration, and the fee-sharing app store. Several are Finn's. | finnbags · ramyodev | open |
| IDX | indexer #153 MetaDAO/Futarchy · #152 Candle · #151 stock-quote catalog | Launchpad + quote coverage continues: MetaDAO + Futarchy, Candle launchpad + trades, a stock-quote catalog refresh. The coverage farm never really stops. | unbound-bags · alaabags | open |
| WEEK 8 | bagsapp #661 · Top Traders tab + trader profile | +736 — the leaderboard storefront, now in its eighth week. The PnL engine behind it shipped Datadog charts weeks ago. Every deck since Aug 3 has said ship or close. | finnbags | created Aug 3 |
| STALE | bagsapp #675 fomo feed + grok chat · #671 discovery refresh | +2,752 / +2,309 — Finn's mobile discovery/feed ideas, still idle. If the RH v3 / BagsPay web work is the new surface, decide whether these rebase or close. | finnbags | idle |
| STALE | bagsapp #649 Cashback + Hyperliquid perps (+25,618) · #618/#609/#602 | #649 is 96+ days old and the largest open PR in the org (Cashback home + Hyperliquid perps). The #618/#609/#602 trio dates to 2025. This is dead weight — merge a slice or close. | finnbags · danbagsfm · ronickg | 96d – 14 mo |
backend #1302 creator claims v0 lookup tables (ramyo) · tc #533 claim fix + #532 versioned creator claim txs (ramyo) · automations #134 CI stale-image guard (ramyo) · indexer #116 mongo backfill wiring (alaa) · automations renovate deps (#131/#132/#133). The creator-claim fix is the one that mattered — it unblocked fee withdrawals same-day. Monday-only; the week's merge list will grow.
The queue is top-heavy with mega-PRs. Five open PRs exceed +7K lines (two exceed +19K, one +55K, one +115K docs). Large PRs review slowly and merge riskily — the native-quote fee rework (#142/#1280/#145) is critical and enormous; sequence it deliberately. #649 at 96 days is still the org's largest open PR, and #661 hit week eight. Standing rule holds: nothing over +5K sits more than two weeks without a merge plan or a split.
Monday-only snapshot. bags.automations led because the creator-claim fix rode out through a stack of image bumps + a CI guard. The backend + twitter-claim commits are the actual claim-fix logic. This is not the week's shape — it's the first day of it. The heavy repos (indexer, datapi, backend feature work) show their weight in the open PR queue, not Monday's commits.
Linear closed 1 issue Monday (SUPPORT-35). That's fine for a Monday. The real gap: the Sep 29 RHC free-gas cliff has no Linear issue, no owner, and no posture decision — and RH volume already fell 91% ahead of it. The take-rate anomaly (now recurring) also isn't filed. Two revenue-critical calls — "what do we do about Robinhood post-subsidy" and "why does the Solana take table disagree with volume" — are living in this deck and nowhere else. Both belong in Linear with owners this week.
Robinhood collapsed: 142 ETH ≈ $389K this week vs 1,531 ETH last week (−91%), with pool volume — the sustainable part — all but gone. Solana's prior full week came off the stock-pairs spike (−24% to $1.43M) but Monday Sep 21 printed $619K in a single day, its best since the announcement. The picture inverted: for a month Robinhood was ~88% of protocol take; this week it's a rounding error and Solana is carrying the number.
RHC's free-gas incentive ends Sep 29. The standard fear was a post-subsidy drop; instead volume fell 91% before the subsidy even ended. Pool volume — activity in graduated tokens, the durable kind — went from 1,090 ETH to 38 ETH. Launches halved to ~10/day. This isn't the cliff; it's the run-up to it. The posture question is now urgent: do we invest to retain RH activity (RH v3 fee capture, #535), or accept RH becomes a thin tail and re-center on Solana + mobile? A month ago RH was 88% of take. This week it's ~13%.
The platform-take table (fee_share_platform_accumulated_event) shows $16.7K accrued in 2 days this week vs $4.8K for all of last week — an implied ~2.6% Monday take rate against 0.34% the prior week. That is almost certainly batched or lagged accruals landing, not real single-day economics. This is the same anomaly flagged on Sep 14. The native-quote fee rework (#142/#1280/#145) in review should reconcile it. Ask: eng confirms the accrual timing before anyone quotes a Solana take rate.
Net read: founder sentiment warm, company sentiment mixed-to-soft, and a clear strategic threat forming. The runners ($SALVOR, agentlayer, $PPW) prove the platform still makes coins go — that's the real asset. But competitors are annexing the "creator fees" narrative Bags pioneered: Pump.fun's fee-share has users running scoreboards against Bags, and UsePaid is selling itself as "the Bags that listened." Meanwhile the on-chain reality (RH −91%, Solana take opaque) doesn't match "we are so back." The move isn't louder confidence — it's a receipt: the $619K Monday, the real runners, and creators-can-claim-again, posted with numbers, so the founder voice and the chain agree.
RH volume fell 91% this week with free gas ending in 8 days. Two honest paths: (a) invest to retain — land tc #535 (RH v3 stock-pair + trade fee) so we actually capture value on RH curves and give creators a reason to stay post-subsidy; or (b) accept the tail — declare RH a thin secondary chain, stop spending net-new eng on it, and re-center on Solana + mobile. What we can't do is drift into Sep 29 with no decision. A month ago RH was 88% of take; this week ~13%. Pick a posture and staff it.
The platform-take table shows $16.7K in 2 days this week vs $4.8K all last week — an implied ~2.6% Monday take that can't be real. It's the same anomaly as Sep 14. Either accruals batch/lag (timing artifact) or the DAMM v2 / stock-pair fee path lands somewhere the table double-counts or misses. The native-quote fee rework (#142/#1280/#145) is in review and should fix the accounting — land it and validate against a known week. Until then, nobody quotes a Solana take rate externally or internally.
indexer #142 (+43K), backend #1280 (+55K), datapi #145 (+9.7K) are the fee-accounting overhaul the whole take-rate question depends on. It's critical and enormous — the exact combination that merges badly if rushed. Give it a review owner, a merge order (indexer facts → backend utilities → datapi analytics), and a validation step against last week's numbers before it touches prod fee math. Don't let it sit unreviewed either — it's the unlock for Obj 2.
Competitors are annexing "creator fees" — Pump.fun users run scoreboards against our 6K number, UsePaid sells itself as "the Bags that listened." The counter isn't "we are so back"; it's proof: this week's real runners ($SALVOR +22,500%, $PPW, agentlayer), the $619K Monday, and "creators can claim again" (we shipped the fix same-day). One thread: real coins, real fees paid to creators, claim bug fixed in a day. Let the founder voice and the chain say the same thing.
#661 Top Traders: week eight — ship it in the next release or close it; this is the fourth deck saying so. #649 Cashback (+25.6K, 96 days): merge a slice or close — it's the org's largest zombie. #675/#671: rebase onto the current surface or close before they rot further. Treasury tracker cron: fix it or the card comes off the deck. The scanner: the stale-cache bug recurred and burned pipeline time twice now — fix `week-scan2.ts` to auto-refresh the cache (or merge the two scanners), so a third deck doesn't ship a false "0 launches."
Read: Linear closed 1 issue Monday and the two most consequential decisions this week (Robinhood posture, take-rate reconcile) exist only in this deck. Tracking discipline hasn't improved — but this week the stakes are higher because both are revenue-defining calls.
Not auto-created — say the word and I'll file these with teams + priorities.
This is a week to make the founder voice and the chain agree. "We are so back" is fine as a vibe; it needs receipts under it — the $619K Monday, the real runners, creators-can-claim-again. And it's the week to answer the creator-fee attack before it hardens into consensus. Rule: every @BagsApp post carries a real number. Hold the Robinhood victory-lap content until the Sep 29 posture is decided — don't hype a chain that's down 91%.
Cadence rule this week: TUE (claims fix) and WED (Solana Monday) are non-negotiable — they're the two receipts that answer the two biggest bear threads. And a hard rule: no Robinhood hype until the Sep 29 posture is decided. Hyping a chain that's down 91% is how "we are so back" becomes a meme against us.
Goal: $100M cumulative protocol revenue. Banked all-time: $32.07M (DefiLlama — Solana basis; Robinhood revenue is additive on top and not captured here). We're 32.1% there. The week added +$26.6K public (Solana-only, −52% WoW) plus ≈$3.9K RH protocol half. The honest read of the number this week: the Robinhood contribution just fell off a cliff, so the blended run-rate compressed hard — and the only thing holding the line is Solana's Monday spike. The measurement chasm is back open because DefiLlama can't see Robinhood; the reconciliation depends on the fee rework landing.
| Milestone | Cumulative revenue | Increment | Status |
|---|---|---|---|
| Today (Sep 21) MEASURED | $32.07M | +$26.6K last week (Solana public) | 32.1% of goal |
| Week close (Sep 27) PACE | ~$32.10M | ~+$30K (compressed w/ RH down) | RH ≈ 13% of take, falling |
| End of Q3'26 (Sep 30) TARGET | $36M | +$3.9M in 9 days | not happening — ~$3M/wk gap. Missed. |
| End of Q4'26 TARGET | $61M | +$29M | needs ~$2.2M/wk from Oct 1 |
| $100M GOAL | $100M | +$68M | H1'27 at a recovered pace |
Basis: DefiLlama protocol revenue, Solana (all-time $32.07M, fees $64.16M). The honest read: Q3's $36M is gone — 9 days, $3.9M, no chance — and the deck says so plainly. What changed this week is the mix: Robinhood, which was 88% of take a month ago, cratered to ~13% and falling into its Sep 29 cliff, so the blended number now leans on Solana. The two levers that change the slope are the same two agenda decisions — the Robinhood posture and reconciling/landing the fee accounting — plus the Solana Monday proving the platform can still print when a narrative catches. This is a floor-finding week, not a growth week.
bags_trade_feed, UTC days. Prior full week Sep 14–21 = $1.433M (−24% off the $1.90M stock-pairs week). Mon Sep 21 = $619K — best single day in a fortnight (1,887 traders, 21 launches). The floor held ~$170K; the Monday pop is the signal.
Corrected on-chain scan (cache refreshed to block 69.43M, +77 tokens). Week 142.4 ETH ≈ $389K, 69 launches, fees ≈ $7.8K. Prior week (corrected) 1,531 ETH / 223 launches → −91% / −69%. Curve 104 ETH / pool 38 ETH — pool activity nearly gone. Free gas ends Sep 29.