task_210726_5 — Platform Tokenomics & Fee Infrastructure (Robbert)

[umeradl]

cc @robertocarlous
Task assigned for Jul 21–27, 2026 (5 working days; deadline Monday, July 27). Full spec: Developer/tasks/task_210726_5.md

The task file has the full specification — exact contract code, function signatures, deployment/wiring order, and test plan. This post is just the summary and assignment notice.

Summary

Issues #37 (staking), #42 (tokenomics), and #43 (fees & treasury) are now design-resolved, and the underlying contract work is genuinely coupled — a new DinTreasury and DinFeeRouter are shared infrastructure both tokenomics and fees need. Combined into one 5-day task rather than three separate ones:

New: DinTreasury.sol — pure holding/spend contract for ETH and ERC20, no split logic, deploys first (no dependencies).

New: DinFeeRouter.sol — owns all fee-split logic. Separate DIN and ETH split configs; the ETH config has no burn bucket at all — structurally, not just by convention, since native ETH can't be burned. Routes DIN protocol fees (burn/treasury/validatorPool/storage/publicGoods) and ETH network fees (treasury/validatorPool/storage/publicGoods, no burn).

DinToken.sol — adds a plain burn(uint256) self-burn function.

DinCoordinator.sol — adds a settable mint cap, a one-way faucet-retirement switch, and reroutes deposited ETH to DinTreasury instead of owner().

DinValidatorStake.sol — converts MIN_STAKE/UNBONDING_PERIOD from constants to owner-settable storage, adds (currently-unenforced) per-model stake bounds storage, and makes jailing real — jailValidator/reactivate — since the Jailed status has existed as dead code until now.

DINModelRegistry.sol — adds a DIN-denominated fee path alongside the existing ETH fees (not replacing them yet), routed through DinFeeRouter.

Scope is platform-contracts only for this task — task-level contract enforcement (DINTaskCoordinator/DINTaskAuditor), DinEmission, delegation, and tombstoning are all explicitly out of scope; see the task file's "Scope boundaries" section for why.

Reference material

• Developer/design/MECHANISM_DESIGN.md §7–§9
• Developer/design/staking-design.md
• Developer/design/tokenomics-design.md
• P3 staking: DAO-settable floor, per-model stake requirements, concurrent-registration cap, withdrawal queue #37, P3 tokenomics: token utility, emission schedule, supply policy, treasury (P3-5.1/5.2) #42, P3 fees & treasury: fee lines, fee-router contract, DIN denomination (P3-5.3) #43 — all updated with the resolved decisions this task implements

If anything in the task spec doesn't feel right, is ambiguous, or needs clarification — comment here and cc @umeradl.

[robertocarlous]

Hi Umer.

I will look into this. Thank you

[umeradl]

Reviewed #51, #60, and #65 together against this task's spec (Developer/tasks/task_210726_5.md). Individual review comments are on each PR; cross-PR findings here since they only show up when you look at all three at once.

Issue coverage

Issue	Delivered by	Status
#43 (fees & treasury)	#51	Done, clean — 112/112 tests, two review rounds both resolved, no blockers.
#37 (staking)	#65	Done, clean — logic correct, 66/66 tests, no blockers.
#42 (tokenomics)	#60 only	Mint-cap/faucet-retirement and the 50/50 slash burn/treasury split are sound in isolation, but bundled inside a PR that also carries stale duplicate copies of #51's and #65's scope — one of those copies has a regression the originals don't (see #60's review).

Every deliverable in the spec's checklist genuinely exists in code somewhere across the three PRs — including the split-sum fuzz test and no-ETH-burn invariant test the spec called out as highest-value. The gaps are about how the three PRs relate to each other, not missing functionality.

Files touched, by PR (head commits: #51 d23a2c9, #60 b8986a4, #65 2787e9f):

File	#51 (d23a2c9)	#60 (b8986a4)	#65 (2787e9f)
foundry/script/DeployPlatform.s.sol	✓	✓
foundry/src/DINModelRegistry.sol	✓	✓
foundry/src/DinCoordinator.sol	✓	✓
foundry/src/DinFeeRouter.sol (new)	✓	✓
foundry/src/DinToken.sol	✓	✓
foundry/src/DinTreasury.sol (new)	✓	✓
foundry/src/DinValidatorStake.sol		✓	✓
foundry/test/DINModelRegistry.t.sol	✓	✓
foundry/test/DeployPlatform.t.sol	✓	✓
foundry/test/DinCoordinator.t.sol	✓	✓
foundry/test/DinFeeRouter.t.sol (new)	✓	✓
foundry/test/DinToken.t.sol	✓	✓
foundry/test/DinTreasury.t.sol (new)	✓	✓
foundry/test/DinValidatorStake.t.sol (new)		✓	✓
foundry/test/SecurityFindings.t.sol	✓	✓
foundry/test/UpgradeValidation.t.sol	✓	✓
Developer/rejected-ideas/stake-weighted-validator-selection.md (new)			✓
Unique to this PR	0	0	1

The starkest single number here: #60 touches 16 files and none of them are uniquely its own — 14 overlap #51, the other 2 overlap #65. What's actually unique to #60 is specific lines within DinCoordinator.sol and DinValidatorStake.sol, not any whole file.

Blocking items (detail in each PR's own review comment):

1. Feat/p3 tokenomics #60 — minStake() accessor removed, breaks DINTaskAuditor/DINTaskCoordinator slashing end-to-end. Confirmed via a real test regression, not theoretical.
2. Feat/p3 tokenomics #60 — DINModelRegistry.sol is missing the FeeRouterNotSet guard Feat/p3 fee router #51 already added in its 07-29 follow-up; would be reintroduced if Feat/p3 tokenomics #60 merges over Feat/p3 fee router #51 unrebased.

Non-blocking flag, worth documenting later: #60 and #65 both reorder DinValidatorStake's storage layout — new governable-param vars land before the existing slasherContracts/validators mappings instead of appended before __gap (confirmed via forge inspect against the develop baseline). Since neither is deployed yet, this is still a genesis proxy either way, not an upgrade — no live layout to break, so no action needed now. Once one of these actually deploys, that becomes the genesis layout every future upgrade has to append against — worth a doc note at that point, not a code change today.

Merge-order recommendation

Splitting the combined task into three narrower PRs was a fine call — no issue with that itself. But right now #60 (opened 07-23, untouched since) is carrying full stale copies of #51 and #65's file scope rather than just its own net-new slice. Recommended path:

1. Merge Feat/p3 fee router #51 as-is.
2. Merge Feat/p3 staking #65 — no blockers, this becomes the canonical DinValidatorStake.sol.
3. Rebase Feat/p3 tokenomics #60 down to just its net-new commits (mint cap/faucet retirement + the burn/treasury split, reapplied onto Feat/p3 staking #65's DinValidatorStake.sol) — that's issue P3 tokenomics: token utility, emission schedule, supply policy, treasury (P3-5.1/5.2) #42's actual unique scope and currently the only place any of it exists. This also avoids re-reviewing DinFeeRouter.sol/DinTreasury.sol a third time, since they're byte-identical to Feat/p3 fee router #51's.

Happy to help with the #60 rebase once #51/#65 land, if useful.

[robertocarlous]

Hi @umeradl

I have fixed the issue

fix(DinValidatorStake): restore minStake() view accessor — added back minStake() external view returns (uint256) which task contracts call to check the stake floor. The function was lost when MIN_STAKE was converted from a constant to a storage variable.

fix(DINModelRegistry): guard DIN entrypoints with FeeRouterNotSet check — both requestModelRegistrationDIN and requestManifestUpdateDIN now revert with FeeRouterNotSet if the fee router hasn't been set yet, matching the guard that PR #51 added.

[umeradl]

Closing: verified against develop (d18f82a)

Re-checked the 2026-07-31 cross-PR review and follow-up fix comment above against current develop. Everything is landed.

PRs merged, in the recommended order:

• Feat/p3 fee router #51 (fee router, issue P3 fees & treasury: fee lines, fee-router contract, DIN denomination (P3-5.3) #43) → d6c6243, later scoped ETH-only per Abraham's fee-denomination direction → 06190e8
• Feat/p3 staking #65 (staking, issue P3 staking: DAO-settable floor, per-model stake requirements, concurrent-registration cap, withdrawal queue #37) → f934a7c, reconciled → bcc0953
• Feat/p3 tokenomics #60 (tokenomics, issue P3 tokenomics: token utility, emission schedule, supply policy, treasury (P3-5.1/5.2) #42) → 2ccc204, reconciled → 77395df

Both blocking items from the review, fixed on develop:

1. minStake() view accessor restored — foundry/src/DinValidatorStake.sol:534.
2. FeeRouterNotSet guard present on both DINModelRegistry.sweepFeesToRouter() and DinCoordinator.sweepFeesToRouter(). Note: it's no longer on requestModelRegistrationDIN/requestManifestUpdateDIN specifically, because those DIN-denominated entrypoints were later removed entirely (06190e8) when the fee-denomination direction moved to ETH-only protocol fees — a separate, later, intentional decision, not a regression of this fix.

Rest of the task spec checks out too: jailValidator/reactivate are real (DinValidatorStake.sol), DinToken.burn(), DinCoordinator's mint cap + one-way faucet retirement, DinFeeRouter.sol/DinTreasury.sol with tests, and the #65-unique Developer/rejected-ideas/stake-weighted-validator-selection.md are all present. The non-blocking storage-layout flag from the review remains correctly deferred (doc note only, once something actually deploys) — no live proxy yet, so no action needed.

Closing — task_210726_5 is fully delivered on develop.