Claim & publish semantics

Normative from the recovered templates [src]; system-level behavior verified onchain [chain].

Across generations, “claim” always means the same physical operation — surrender child tokens, receive locked parent collateral — but the gating changes per generation:

Generation Claim shape Gate Publish
V1 none — permanent sink — —
V2 Claim(Contract, Amount) — open sibling: any registered sibling with Debenture() == true is a valid key key must be an unpublished registered sibling publish() (level ≥ 100) sets Debenture = false: stops being a key, enables withdraw()
V3 Claim(Contract, Amount) — gated sibling key must have the same Creator and same Parent as the payout token; payout token must still have Debenture == true; payout token’s parent must not be TBILL (the V1 sink stays a sink) none on the token — the minter’s TTI interface declares publish() but the deployed template implements it nowhere; first withdraw() sets Debenture = false — the payout side dies on first exit
V4 Claim(Amount) — self-redeem burn your own tokens; parent pays out; TBILL-parented children excluded (same guard as V3) n/a (no Debenture flag)

One asymmetry is easy to miss and matters for anyone holding V2 keys: the TBILL-parent guard (if (Parent == TreasuryBillContract) revert) exists in the V3 and V4 templates but not in V2 [src] — a V2 child of TBILL remains fully claimable, and locked TBILL can leave through it. The /claim tool warns on exactly this case.

Two system-level consequences, verified onchain [chain]:

  1. Publish ≠ protection. Publishing removes a token as a claim key; it does nothing to stop claims on a token as the payout source. The bureau-ancestry spine FDIC → DFM 0x51160F352ED148C89d48dfe6384Edd07aFA24E0E → PARADE 0xE37ACc54711562510FaFC45d8199Ee329ebBceDd → Tellerz was published early in its life — and was then claim-loop-farmed for ~20 months, still active as of 2026-09. Publishing is an ownership gesture, not a defense.
  2. Unpublished V2 children remain permanent claim keys by design. 511 V2 children were ever published (event-study scan [chain]); the unpublished majority are keys forever (their entity pages show the backing ratio that follows from being farmed or not, and the generated panel on the V2 page carries the current publish-adoption counts).

The economics that made this loop profitable: a claim pays out parent collateral 1:1 against surrendered child supply — where child supply was minted 1:1 (V2) — so any key whose market price falls below its parent value is farmed until the parent custody hits zero (visible on entity pages as the 0%-backing warning).