Point the delegation-copy notes at the authorization voice

TODO.md

@@ -22,7 +22,7 @@ Getting **testnet to a two-person shared lab** is also a standing plan, not a pi
 
 - **(Tell)** Send the donor-set waiting-period page. Note and project screens show a pending spend's amount and deadline as money that can still be cancelled. `shouldPageDonor` still only decides once per `(noteId, nonce)`. There is no opt-in store and no email or push sender. Public remarks from people who are not flaggers stay a later UI feature and are not stored on-chain. Rules: [waiting-period.md](specs/tech/subsystems/delegation/waiting-period.md).
 
-- One voice for delegation copy. The donor is authorizing an address to spend a stated amount on projects in Commonality. The delegate promises nothing. A stated intent is public and does not bind the spend. Unspent funds stay revocable by the donor. Commonality does not hold the funds, choose the delegate, or supervise the spending. Remove the steward voice: entrusting money to a scout, program-officer framing, "money under management," and any Commonality ranking whose job is to send people to a delegate. A public history of what an address already funded can stay. "Scout" as the early contributor who may later be reimbursed at cost can stay; do not let that word mean a manager of other people's money. Start with `specs/product/legal/retroactive-funding-redesign.md` (Design 2) and `docs/end-user/lazyGiving/` (`retroactive-funding.md`, `index.md`, `fund-something.md`, `get-your-project-funded.md`). No delegate marketplace.
+- One voice for delegation copy. Not started, and not a contract change. Write toward the baseline in [delegation-narrowing.md](specs/product/legal/delegation-narrowing.md): the donor authorizes an address to spend a stated amount on projects in Commonality; the delegate promises nothing; a stated intent is public and does not bind the spend; unspent funds stay revocable by the donor; Commonality does not hold the funds, choose the delegate, or supervise the spending. Do not describe the delegate as a steward. "Scout" stays an early contributor who may be reimbursed at cost, not a manager of other people's money. A public history of what an address already funded can stay. No delegate marketplace, and no Commonality ranking whose job is to send donors to a delegate. Design 2 in [retroactive-funding-redesign.md](specs/product/legal/retroactive-funding-redesign.md) is the old framing to correct (program officer, money under management), not the target. The same voice is still in the rollout tracker, the "delegated budget" reward line in [securities.md](specs/product/legal/securities.md), `docs/end-user/lazyGiving/` (`retroactive-funding.md`, `index.md`, `fund-something.md`, `get-your-project-funded.md`), [become-a-delegate.md](docs/end-user/alignment/become-a-delegate.md), and the Suggested delegates panel in `ui/src/fundingportals/components/SuccessfulProjectsList.tsx`.
 
 
 - **(Tell)** Testnet Commonality SPA does not hydrate. `https://testnet.commonality.works/` and deep links now return the HTML shell (SPA fallback after public-gateway 429 is deployed), but browser loads fail on chunks such as `/assets/address-TZjglcQ5.js` (HTTP 429, public IPFS sunset body). Dedicated Pinata origin times out; Worker then falls through to `ipfs.io` / `w3s.link`. Reproduce, fix the Worker/gateway path so real assets are served from Pinata (or another working origin) instead of caching/returning 429, redeploy `cloudflare-ui-gateway`, and verify `/`, `/founders`, and a hydrated heading in a real browser. Pinata dashboard Host Origins remains Adam’s step in [`inbox.md`](inbox.md). Continuity: [`continuity/2026-09-15-commonality-live-gateway-followup.md`](continuity/2026-09-15-commonality-live-gateway-followup.md).

specs/product/legal/retroactive-funding-redesign.md

@@ -14,9 +14,11 @@ directory: a map for where to spend lawyer money, not legal advice.)
 
 ## Resolved decisions (Jul 2026)
 
-Adam has chosen this redesign (Design 1 + Design 2 below) as the securities
-posture to build toward, and settled the open design questions. These are the
-decisions of record; the analysis further down explains *why* each is safe.
+Adam chose Design 1, below, as the reimbursement posture, and settled the
+open design questions in that section. Those are the decisions of record; the
+analysis further down explains *why* each is safe. Design 2 was the Jul 2026
+complement. Its steward framing is not the current voice — see the note on
+that section.
 
 - **Reimbursement is pro-rata and simultaneous, with no seniority.** Early
   contributors all funded before the assurance threshold was met, so they are
@@ -150,6 +152,14 @@ irrational anyway, so the capped market was mostly theater.
 
 ## Design 2 (complement): Delegation is the legal scout-payment mechanism
 
+**Superseded for the voice (2026-09).** This section records the Jul 2026
+complement: delegation as the scout's reward, framed as program-officer work
+and money under management. That is no longer the intended description. The
+live baseline is [delegation-narrowing.md](delegation-narrowing.md): the donor
+authorizes a spend, the delegate promises nothing, and "scout" stays an early
+contributor who may be reimbursed at cost. The copy scrub is still open in
+[TODO.md](/TODO.md). The paragraphs below are the framing that was chosen then.
+
 We already built the clean version of "skilled people direct other people's
 money": the **Delegation** subsystem. Reframed:
 

specs/product/legal/securities.md

@@ -45,7 +45,7 @@ This arguably dominates the donation-first posture, since it keeps retroactive f
 
 ### Chosen posture: reimbursement waterfall, no market (Jul 2026)
 
-One step past the contract-level cap: make receipt tokens non-transferable and delete the secondary market entirely; retroactive donations route through an on-chain waterfall (pro-rata scout reimbursement at cost, overflow to the project), and the scout reward moves to reputation/delegated budget. This removes the exchange/dealer category outright rather than shrinking it, and structurally forecloses the third-party markup-venue scenario below. Full analysis: [retroactive-funding-redesign.md](retroactive-funding-redesign.md).
+One step past the contract-level cap: make receipt tokens non-transferable and delete the secondary market entirely; retroactive donations route through an on-chain waterfall (pro-rata scout reimbursement at cost, overflow to the project), and the scout reward is the public reimbursement record, not a larger budget to manage. Treating delegation as that payment channel was the Jul 2026 complement; that voice is superseded by [delegation-narrowing.md](delegation-narrowing.md). This removes the exchange/dealer category outright rather than shrinking it, and structurally forecloses the third-party markup-venue scenario below. Full analysis of the waterfall: [retroactive-funding-redesign.md](retroactive-funding-redesign.md).
 
 **What if a genuine third party builds an uncapped market/UI?** Mostly their problem, not ours — and *Gary Plastic* itself is the authority: the banks issuing the CDs had no securities problem; Merrill, the third party that built the market and marketed the flip, did. Issuer and market-promoter are analyzed separately on their own conduct. Our offering is judged by what *we* led purchasers to expect at the time of sale (third-party appreciation is the collectibles situation — scalped concert tickets don't become securities). The third party running a markup venue may be creating *their own* investment-contract offering — their exposure, worth warning would-be community-UI operators about explicitly. Our protection depends on the separation staying real: (1) **no integration** — no DelegatableNotes adapter for the markup venue, no displaying its listings/prices on our surfaces, no linking to it; (2) **no profit from it** — in particular no ERC-1155 royalty hooks collecting on third-party markup sales; (3) **no promotion** — one "you can get more elsewhere" wink reattaches the expectation to us, and the *old* profit-narrative docs are historical evidence of intent, so the scrub matters under every posture; (4) **no theater** — if markup trading becomes the dominant known use and our story stays unchanged, substance-over-form can eventually reach us. Caveat: the SEC token framework lists secondary-market availability as a profit-expectation factor, so a third-party market is *a* factor against us — a weak one when everything we control says reimbursement-only. Design implication: this strengthens the contract-level cap over the UI-level cap — with a UI cap, a "third-party UI" is just another front-end on *our* market contract and markup listings land in our event stream for our own indexer/UIs to filter; with a contract cap, a third party must deploy a structurally separate market, making the separation legible and unmistakably theirs.
 

workflow/retroactive-funding-redesign-rollout.md

@@ -23,9 +23,12 @@ while keeping retroactive funding:
   in** — never interest/premium/markup. The secondary market is **deleted
   entirely** (a capped market is vestigial; removing it eliminates the
   exchange/dealer problem instead of merely shrinking it).
-- **Design 2 — Delegation is the scout-payment channel.** Reward for judgment
-  is reputation + larger *delegated budgets* (money to manage, never in
-  pocket), optionally a flat service fee — never per-dollar-of-outcome.
+- **Design 2 — Delegation is the scout-payment channel.** Jul 2026 complement,
+  superseded for the voice. It rewarded judgment with larger delegated budgets
+  (money to manage, never in pocket). The current baseline is the one-voice
+  note in [delegation-narrowing.md](/specs/product/legal/delegation-narrowing.md):
+  a limited authorization, not a steward. The copy scrub is still open in
+  [TODO.md](/TODO.md). Do not implement this bullet as written.
 
 The marketing/UX invariant that governs everything below: **"get your money
 back and fund the next one" must be the *whole* story.** No buy/sell/invest/