feat(indexer): wire P3-staking events — ValidatorJailed/Reactivated, governable MIN_STAKE, JailEvent entity

[robertocarlous]

Summary

Follow-up to PR #29 (feat/din-indexer) addressing the P3-staking gap identified in the cross-PR review (discussion #47). Branches off feat/din-indexer and merges feat/p3-staking (PR #65) locally so the subgraph stays in sync with the staking contract additions without blocking on #65's own merge timeline into develop.

Root cause: feat/din-indexer was written before PR #65 added six new events to DinValidatorStake and converted MIN_STAKE from a Solidity constant to a governable storage variable. The event-coverage-audit (§2.2) flagged this exact scenario: "when a jailing mechanism is added, it must emit a ValidatorJailed event with the jailedUntil timestamp for the indexer to reconstruct status correctly."

What changed

• abis/DinValidatorStake.json — regenerated from the compiled P3 contracts; includes all new functions, events, and errors
• subgraph.yaml — registers handlers for all six new events: ValidatorJailed, ValidatorReactivated, MinStakeUpdated, UnbondingPeriodUpdated, ModelStakeBoundsUpdated, MaxConcurrentRegistrationsPerStakeUnitUpdated
• src/staking.ts — adds the six corresponding handler functions; introduces a StakeGovernance singleton entity for governance-parameter writes and ModelStakeBounds entity writes
• src/utils.ts — syncValidatorStatus gains a Jailed branch (jailedUntil != null → status "Jailed"); reads the current floor from StakeGovernance via getMinStake() instead of a hardcoded constant — no silent auto-return from Jailed, mirroring _syncValidatorStatus() semantics
• schema.graphql — adds jailedUntil field and jailEvents relation to Validator; new JailEvent entity; new StakeGovernance singleton type; new ModelStakeBounds type

Also included from the parent branch (feat/din-indexer-65 base commits):

• subgraph/scripts/update-subgraph-addresses.sh — patches subgraph.yaml placeholder addresses from a deployments.json (item Bump web3 from 7.10.0 to 7.15.0 in /dincli in the pip group across 1 directory #3 from review)
• Known-limit comments on both first: 1000 GraphQL queries in dincli/cli/dindao.py (item Bump cryptography from 46.0.4 to 46.0.7 in /dincli in the pip group across 1 directory #2 from review)

Merge order

This PR should merge after PR #29 (feat/din-indexer) and after PR #65 (feat/p3-staking) both land. It is not a blocker for either.

[robertocarlous]

Hi @umeradl

Pls can you look into this. thanks

[umermjd11]

Review — against 1a2a877 (PR head), merge-base 866521a (feat/din-indexer / PR #29 HEAD)

Scope confirmed: this PR's base is PR #29 (feat/din-indexer), not develop. git diff 866521a..1a2a877 — the actual work this PR itself contributes — touches subgraph/abis/DinValidatorStake.json, subgraph/schema.graphql, subgraph/scripts/update-subgraph-addresses.sh, subgraph/src/staking.ts, subgraph/src/utils.ts, subgraph/subgraph.yaml, and dincli/cli/dindao.py (comments only) — matches the PR body's own "What changed" list exactly. Everything else in the raw diff (foundry/hardhat P3-staking contracts, their tests, ~30 doc files) is inherited unchanged from this PR's local merge of already-merged PR #65 (feat/p3-staking) into the feat/din-indexer line, not new work from this PR's own commits.

Files changed (this PR's own contribution)

File	Change	What it does
subgraph/subgraph.yaml	+15	Registers handlers for the 6 new DinValidatorStake events (ValidatorJailed, ValidatorReactivated, MinStakeUpdated, UnbondingPeriodUpdated, ModelStakeBoundsUpdated, MaxConcurrentRegistrationsPerStakeUnitUpdated) and 3 new schema entities.
subgraph/src/staking.ts	+97	Adds the 6 handler functions; introduces loadOrCreateStakeGovernance() singleton pattern for governance-parameter writes.
subgraph/src/utils.ts	+29/-modified	syncValidatorStatus gains a Jailed branch; getMinStake() reads the governable floor from StakeGovernance instead of a hardcoded constant.
subgraph/schema.graphql	+73	Adds jailedUntil/jailEvents to Validator; new JailEvent, StakeGovernance, ModelStakeBounds types.
subgraph/abis/DinValidatorStake.json	regenerated	Includes the new P3-staking functions/events/errors.
subgraph/scripts/update-subgraph-addresses.sh (new)	+78	Patches placeholder addresses in subgraph.yaml from a deployments JSON.
dincli/cli/dindao.py	+6	Comments documenting the known first: 1000 GraphQL result cap.

Independently re-verified

• All 6 event handlers, read directly against the actual event signatures in subgraph.yaml/the regenerated ABI: correct field mappings, no signature mismatches.
• syncValidatorStatus: correctly implements "Jailed status has no silent auto-return — stays Jailed until reactivate() explicitly clears it." I checked this against develop's actual, currently-merged DinValidatorStake._syncValidatorStatus() (not this branch's own embedded copy — see finding below) and it's a precise match: if (validator.status == ValidatorStatus.Jailed) { return; }, no time-based auto-clear. Good catch by whoever wrote this — it would have been an easy place to get subtly wrong.
• handleValidatorSlashed: mirrors slash()'s active-stake-then-pending-withdrawal consumption order exactly, including the pendingWithdrawal zeroing edge case.
• Governance singleton pattern (StakeGovernance, ModelStakeBounds): sensible design — seeds initialize() defaults on first load, updates in place on each governance event, matches the "inert storage, not yet enforced" framing from the contract's own NatSpec.
• npx graph codegen: clean, generates types for all 6 data sources with no errors.

Real finding: this PR's embedded DinValidatorStake.sol has drifted from what's actually on develop, and the indexer is missing a governance event as a direct result

This PR locally merges feat/p3-staking (PR #65) rather than waiting for it to land on develop — reasonable, since #65 had already merged separately by the time this PR was opened. But the two copies have diverged. Diffing this branch's foundry/src/DinValidatorStake.sol against develop's real, currently-merged version: develop has a slashTreasury mechanism this PR's copy doesn't — IBurnableToken interface, slashTreasury state var, SlashTreasuryUpdated event, setSlashTreasury() setter, and a 50%-burn/50%-treasury split inside slash(). This isn't in the version #65 was merged in with when this branch was cut.

Concrete, checkable consequence: SlashTreasuryUpdated and slashTreasury appear nowhere in this PR — not in subgraph.yaml, not in staking.ts, not as a field on the StakeGovernance entity alongside minStake/unbondingPeriod/maxConcurrentRegistrationsPerStakeUnit (which is exactly the entity built to hold DAO-governable params like this one). Not a bug in the code this PR wrote — a real gap caused by the snapshot it was built from being stale, and invisible to GitHub's conflict detection since this PR targets feat/din-indexer, never touches develop directly, and the two versions don't textually conflict (they just silently diverged).

Real finding: npx graph build fails — but it's pre-existing in PR #29, not introduced here

subgraph/src/registry.ts doesn't compile: AssemblyScript doesn't support TypeScript union-type parameters, and two helper functions there take event: ModelDisabled | ModelEnabled / event: OpenSourceFeeUpdated | ProprietaryFeeUpdated | OpenSourceUpdateFeeUpdated | ProprietaryUpdateFeeUpdated. There's also a harmless-but-also-broken inline import("@graphprotocol/graph-ts").Address type reference in the same file (redundant — Address is already statically imported at the top).

Confirmed via direct diff that registry.ts is byte-identical between this PR and PR #29's actual current head — this is entirely inherited, not something this PR's own commits touch or introduce. But it does mean the subgraph can't currently be built or deployed end-to-end with the real graph-cli toolchain, which blocks verifying this PR's own staking.ts the same way. Reviewed staking.ts/utils.ts by careful static reading against the ABI and contract source instead — no issues found there, and graph codegen (which parses the ABIs/schema independent of the AssemblyScript compile step) succeeds cleanly, giving reasonable confidence the types line up.

Follow-ups

• Add SlashTreasuryUpdated/slashTreasury tracking to StakeGovernance (event handler + schema field) before this indexer work is considered complete for the currently-deployed contract shape.
• Fix registry.ts's union-typed parameters (split into per-event-type functions, or a shared base interface) — blocks graph build/graph deploy entirely today. Belongs in feat(indexer): implement DIN Protocol subgraph — platform + task-level contracts #29, but nothing here can be end-to-end verified until it's fixed either there or as a small unblocking commit riding along with this PR.
• subgraph/docs/event-coverage-audit.md (§ "No ValidatorJailed event") predates both the jailing mechanism and slashTreasury — worth a refresh pass once the above lands, since this PR is precisely the kind of change that audit doc exists to track.

Recommendation

The code this PR itself wrote is correct and well-verified — event wiring, status-sync semantics, and the governance-singleton pattern all check out against the real, current contract behavior. Two things worth resolving before treating the indexer as complete: the slashTreasury gap (concrete, scoped, this PR's responsibility since it's the one claiming to sync staking events) and the registry.ts build blocker (not this PR's bug, but worth linking to #29 so it doesn't get lost). Neither blocks merging into feat/din-indexer today.

[umermjd11]

Files changed — as of 1a2a877 (PR head)

Diffed against merge-base 866521a, which is PR #29's (feat/din-indexer) actual current head — confirmed as a real ancestor of this branch (git merge-base --is-ancestor checks out clean). This PR's own diff against that base: 165 files, 12,899 insertions(+) / 1,780 deletions(-), but the large majority of that is inherited unchanged content from this branch's local merge of already-merged PR #65 (feat/p3-staking), not new work — see the table below for what this PR's own commits actually touch.

(Table is per-file, key/value.)

subgraph/subgraph.yaml

Field	Value
Change	Modified
Lines	+15
Diff (what exactly is in this PR)	Registers the 6 new DinValidatorStake event handlers and 3 new schema entities on the existing DinValidatorStake data source.
Diff vs PR #29 current HEAD	This is the diff — #29 doesn't have these registrations.
Local merge conflict	No
GitHub merge conflict	No

subgraph/src/staking.ts

Field	Value
Change	Modified
Lines	+97
Diff (what exactly is in this PR)	6 new handler functions + loadOrCreateStakeGovernance() singleton helper.
Diff vs PR #29 current HEAD	This is the diff.
Local merge conflict	No
GitHub merge conflict	No

subgraph/src/utils.ts

Field	Value
Change	Modified
Lines	+29 (net)
Diff (what exactly is in this PR)	Jailed branch in syncValidatorStatus; getMinStake() reads from StakeGovernance instead of a constant.
Diff vs PR #29 current HEAD	This is the diff.
Local merge conflict	No
GitHub merge conflict	No

subgraph/schema.graphql

Field	Value
Change	Modified
Lines	+73
Diff (what exactly is in this PR)	jailedUntil/jailEvents on Validator; new JailEvent, StakeGovernance, ModelStakeBounds types.
Diff vs PR #29 current HEAD	This is the diff.
Local merge conflict	No
GitHub merge conflict	No

subgraph/abis/DinValidatorStake.json

Field	Value
Change	Modified (regenerated)
Lines	large (full ABI regen)
Diff (what exactly is in this PR)	Regenerated from the (locally-merged) P3-staking contract compile.
Diff vs PR #29 current HEAD	This is the diff — but see the review comment above: this ABI reflects a stale snapshot of DinValidatorStake.sol that's missing the slashTreasury additions actually on develop.
Local merge conflict	No
GitHub merge conflict	No

subgraph/scripts/update-subgraph-addresses.sh (new)

Field	Value
Change	New
Lines	+78
Diff (what exactly is in this PR)	Patches placeholder contract addresses in subgraph.yaml from a deployments JSON, foundry- or hardhat-schema.
Diff vs PR #29 current HEAD	Net-new file.
Local merge conflict	No
GitHub merge conflict	No

dincli/cli/dindao.py

Field	Value
Change	Modified
Lines	+6
Diff (what exactly is in this PR)	Comment-only: documents the existing first: 1000 GraphQL result cap on two queries.
Diff vs PR #29 current HEAD	This is the diff.
Local merge conflict	No
GitHub merge conflict	No

Everything else in the raw diff is inherited, not new

The remaining ~155 files (foundry/src/DinValidatorStake.sol + its hardhat mirror, foundry/test/DinValidatorStake.t.sol, ~30 Developer/Documentation reorganization files, tests/dincli/*, etc.) come from this branch's local merge commit (ecdd87b) pulling in already-merged PR #65's work, so feat/din-indexer stays buildable against the current staking contract shape without waiting on #65's own merge timeline. None of it is this PR's own authorship — flagged here only because it's where the slashTreasury drift documented in the review comment above actually lives.

No conflicts against the stated base

gh pr view 72: mergeable: MERGEABLE, mergeStateStatus: CLEAN against feat/din-indexer (#29). Confirmed independently: merge-base 866521a is PR #29's real current head, so this PR is sitting directly on top of it with nothing to reconcile.

Downstream risk, not a blocker on merging into #29 today

When #29 (and by extension this PR) eventually merges into develop, foundry/src/DinValidatorStake.sol as carried in this branch will need to reconcile against develop's real version — which has diverged (the slashTreasury mechanism, see review comment). Depending on how that merge is done, this is either a real textual conflict (the safe outcome — forces someone to look at it) or, if resolved carelessly favoring this branch's version, a silent regression that drops slashTreasury from develop. Worth flagging explicitly to whoever handles the eventual #29→develop merge; not something to fix in this PR, which only needs to merge cleanly into #29.

[umermjd11]

Reviewed in an isolated worktree against the real commit graph, not just GitHub's file-change summary.

Scope correction first: GitHub reports 185 files changed / +20937/-1412 against base feat/din-indexer — that's misleading. This PR's head (1a2a877) is confirmed 25 commits ahead of PR #29's head (866521a) and 0 commits behind it, i.e. it fully contains PR #29 (still unmerged into feat/din-indexer). PR #29 is reviewed separately; re-reviewing its content here would be redundant. Of those 25 incremental commits, only 2 are this PR's own work (1a2a877 "wire P3-staking events into subgraph" and 2b015f5 "document 1000-result GraphQL cap; add update-subgraph-addresses.sh"); the rest arrived via ecdd87b merging origin/feat/p3-staking (PR #65, already merged to develop on 2026-08-10) — old history feat/din-indexer-65 absorbed at branch-cut time, not new content. Scoping the diff to subgraph/ (the only directory this PR's own commits actually touch) gives the true content: 6 files, +962/-380 — confirmed via git diff 866521a..1a2a877 --stat -- subgraph/. I also confirmed the other 110 non-subgraph files that differ from current develop are pre-existing drift from develop's 165 commits of forward progress since this branch's snapshot — not anything introduced by this PR.

Claimed: Six new events (ValidatorJailed, ValidatorReactivated, MinStakeUpdated, UnbondingPeriodUpdated, ModelStakeBoundsUpdated, MaxConcurrentRegistrationsPerStakeUnitUpdated) are wired end-to-end.

Verified — all six match exactly: grep'd the live event signatures out of foundry/src/DinValidatorStake.sol (lines 90-95) and diffed them against the event: entries registered in subgraph.yaml (lines 139-150) — indexed-parameter positions match exactly, e.g. ValidatorJailed(indexed address,uint64,indexed bytes32,indexed address) matches the Solidity signature parameter-for-parameter. Each has a corresponding handle* function present in src/staking.ts.

Claimed: abis/DinValidatorStake.json was regenerated from the compiled P3 contracts.

Verified: the checked-in ABI contains all six new events plus the new errors (InvalidJailDuration, JailPeriodNotExpired, NotJailed, InvalidMinStake, InvalidUnbondingPeriod, AmountLessThanMinStake) and functions (jailValidator, setMinStake, setUnbondingPeriod, setModelStakeBounds, setMaxConcurrentRegistrationsPerStakeUnit) — grepped directly out of the JSON, all present.

Claimed: syncValidatorStatus's new Jailed branch mirrors the contract's _syncValidatorStatus() — no silent auto-return once jailedUntil passes; only explicit reactivation clears it.

Verified — exact semantic match: read _syncValidatorStatus() in foundry/src/DinValidatorStake.sol (lines 402-421): once status == Jailed it returns immediately regardless of timestamp, with an explicit comment "No silent auto-return: the operator must prove the node is back online first" — reactivate() (line 358) is the only path that clears it, gated by block.timestamp < v.jailedUntil → revert. The AS mirror in utils.ts checks jailedUntil != null before any other branch and returns early; handleValidatorReactivated in staking.ts sets jailedUntil = null (the only place that does), and handleValidatorJailed is the only place that sets it. Same state machine, same invariant.

Claimed: StakeGovernance singleton seeds correct defaults before any governance event fires.

Verified: loadOrCreateStakeGovernance() seeds minStake = 10000000000000000000 (10 DIN) and unbondingPeriod = 604800 (7 days) — matches the contract's initialize() exactly (MIN_STAKE = 10 * 1e18, UNBONDING_PERIOD = 7 days, foundry/src/DinValidatorStake.sol lines 121-122).

Claimed: ModelStakeBoundsUpdated writes a ModelStakeBounds entity, and the governance updates (MinStakeUpdated/UnbondingPeriodUpdated/MaxConcurrentRegistrationsPerStakeUnitUpdated) all update the same StakeGovernance singleton.

Verified: read all four handlers in staking.ts — all three governance-param handlers call the shared loadOrCreateStakeGovernance() and write to "singleton"; handleModelStakeBoundsUpdated correctly keys a per-model ModelStakeBounds entity by modelId. schema.graphql has matching StakeGovernance/ModelStakeBounds/JailEvent types and the new jailedUntil/jailEvents fields on Validator.

Claimed: subgraph/scripts/update-subgraph-addresses.sh patches subgraph.yaml placeholder addresses from a deployments JSON.

Verified: script present, reads foundry/deployments/localhost.json with a fallback to the hardhat path, patches subgraph.yaml in place. Not executed end-to-end (needs a live local deploy artifact), but the logic reads correctly.

Claimed (PR body, "also included from the parent branch"): known-limit comments on the two first: 1000 GraphQL queries in dincli/cli/dindao.py.

Correction: this file does not appear in this PR's own subgraph/-scoped diff — it's inherited from PR #29's head, not new content in #72. Not this PR's claim to verify; flagging so the merge-proposal table below doesn't imply otherwise.

Verified — codegen/build: npm install && npx graph codegen succeeds cleanly (types generated for all data sources, templates, and schema). npx graph build fails, but the failure is in src/registry.ts (TS1005/TS1110/TS1109 parse errors around line 204/253) — a file this PR does not touch (confirmed absent from the subgraph/-scoped diff). That's a pre-existing AssemblyScript compile issue inherited from PR #29's own work, out of scope here and worth flagging separately on PR #29 rather than blocking this one.

Not independently re-verified: a full green graph build of this PR's own handlers in isolation (blocked by the unrelated registry.ts crash above) and a live local Graph Node index-and-query run (no anvil/graph-node stack running in this worktree).

---

Every checkable claim in this PR's actual content (the 6 subgraph/ files) held up under direct comparison against the live Solidity source, byte-for-byte ABI inspection, and successful codegen. The one build gap (registry.ts) is pre-existing and belongs to PR #29, not this PR. Mergeable as-is once PR #29 lands first — this PR's head already contains #29's commits, so merging #72 alone (without #29 first) would silently pull #29's unreviewed content into feat/din-indexer as a side effect. Recommended order: PR #29 → feat/din-indexer, then PR #72 → feat/din-indexer (fast-forward or trivial merge, since #72 already has #29 fully merged in locally). PR #65's dependency is already satisfied (merged to develop 2026-08-10). Note feat/din-indexer itself is currently 187 commits behind develop — that gap should be addressed when feat/din-indexer merges to develop, not before.

[umermjd11]

Files changed (6) — as of 1a2a877 (PR head)

Scope note: GitHub's own diff stat against base feat/din-indexer (185 files, +20937/-1412) is inflated because this PR's head already contains PR #29's unmerged commits (confirmed: 25 commits ahead of PR #29's head, 0 behind) plus old history absorbed via the feat/p3-staking (PR #65, already merged to develop) merge commit. The table below is scoped to git diff 866521a(PR #29 head)..1a2a877(PR #72 head) --stat -- subgraph/ — the only files this PR's own 2 commits actually touch. GitHub agrees: mergeable: MERGEABLE, mergeStateStatus: CLEAN (against the stale base feat/din-indexer). Local git merge-tree of #72's head against feat/din-indexer + #29 applied first is clean — no conflicts, since #72 already contains #29 verbatim.

(Table format per PR #63's review.)

subgraph/subgraph.yaml

Field	Value
Change	Modified
Lines	+15
Diff (what exactly is in this PR)	Registers 6 new event handlers on the DinValidatorStake data source: ValidatorJailed, ValidatorReactivated, MinStakeUpdated, UnbondingPeriodUpdated, ModelStakeBoundsUpdated, MaxConcurrentRegistrationsPerStakeUnitUpdated.
Functionality — how & why	How: each event:/handler: pair maps a Solidity event signature to a named AS function in staking.ts; the Graph Node indexer calls the handler whenever a matching log is emitted. Why: these six events exist on DinValidatorStake.sol (added by PR #65) but had no subgraph coverage — the event-coverage audit flagged this as a gap since the indexer can't reconstruct jail/governance state without them.
Diff vs current develop HEAD	New relative to feat/din-indexer's stale base; no equivalent subgraph exists on develop yet (this whole directory is still pre-merge).
Recommended merge proposal	Merged as-is — signatures verified byte-for-byte against foundry/src/DinValidatorStake.sol.
Actual merge proposal	Soon
Pending proposal	None
Local merge conflict	No
GitHub merge conflict	No

subgraph/src/staking.ts

Field	Value
Change	Modified
Lines	+97
Diff (what exactly is in this PR)	Adds handleValidatorJailed, handleValidatorReactivated, handleMinStakeUpdated, handleUnbondingPeriodUpdated, handleMaxConcurrentRegistrationsPerStakeUnitUpdated, handleModelStakeBoundsUpdated, plus a loadOrCreateStakeGovernance() helper that seeds the StakeGovernance singleton.
Functionality — how & why	How: jail/reactivate handlers write jailedUntil on Validator and call syncValidatorStatus; a JailEvent audit row is also written per jail. Governance handlers all mutate one StakeGovernance("singleton") entity; ModelStakeBoundsUpdated writes a per-model ModelStakeBounds entity keyed by modelId. Why: mirrors the contract's actual state machine so queries reflect on-chain jail/governance state instead of assuming a hardcoded MIN_STAKE. Verified the singleton's seeded defaults (10 DIN, 7-day unbonding) match initialize() in the Solidity source exactly.
Diff vs current develop HEAD	New file content, no conflicting equivalent on develop.
Recommended merge proposal	Merged as-is — reactivate-clears-jail / no-auto-expiry semantics verified to match _syncValidatorStatus() exactly.
Actual merge proposal	Soon
Pending proposal	None
Local merge conflict	No
GitHub merge conflict	No

subgraph/src/utils.ts

Field	Value
Change	Modified
Lines	+29/-? (net +29 per diffstat)
Diff (what exactly is in this PR)	syncValidatorStatus gains a Jailed branch (checked before Exiting/Active) and now reads the active-stake floor via getMinStake() (pulled from StakeGovernance) instead of a hardcoded constant.
Functionality — how & why	How: getMinStake() loads the singleton and falls back to a seeded default if ungoverned yet; the Jailed check returns early with no further status computation, same as the contract's guard. Why: without this, a jailed validator's status could be silently recomputed to Active/Exiting on the next stake-changing event — the exact bug class the event-coverage audit called out.
Diff vs current develop HEAD	No equivalent on develop.
Recommended merge proposal	Merged as-is.
Actual merge proposal	Soon
Pending proposal	None
Local merge conflict	No
GitHub merge conflict	No

subgraph/schema.graphql

Field	Value
Change	Modified
Lines	+73/-?
Diff (what exactly is in this PR)	Adds jailedUntil: BigInt + jailEvents: [JailEvent!]! @derivedFrom(...) to Validator; new entities JailEvent, StakeGovernance, ModelStakeBounds.
Functionality — how & why	How: @derivedFrom gives Validator.jailEvents a reverse lookup with no extra storage, backed by JailEvent.validator. Why: exposes jail history and governance state to GraphQL consumers (dincli, external dashboards) without requiring separate RPC calls.
Diff vs current develop HEAD	No equivalent on develop.
Recommended merge proposal	Merged as-is.
Actual merge proposal	Soon
Pending proposal	None
Local merge conflict	No
GitHub merge conflict	No

subgraph/abis/DinValidatorStake.json

Field	Value
Change	Modified (regenerated)
Lines	+1050/-380 (largely reformatted/expanded ABI JSON)
Diff (what exactly is in this PR)	Full ABI regeneration from the compiled P3 contract — adds the 6 new events, 6 new errors, and 5 new functions used above.
Functionality — how & why	How: used by graph codegen to generate typed AS event/contract bindings (generated/DinValidatorStake/DinValidatorStake.ts). Why: the handlers in staking.ts can't type-check or compile without the matching ABI entries; confirmed npx graph codegen succeeds cleanly with this ABI.
Diff vs current develop HEAD	No equivalent on develop.
Recommended merge proposal	Merged as-is — spot-checked all new event/error/function names present.
Actual merge proposal	Soon
Pending proposal	None
Local merge conflict	No
GitHub merge conflict	No

subgraph/scripts/update-subgraph-addresses.sh

Field	Value
Change	New
Lines	+78
Diff (what exactly is in this PR)	Shell script patching subgraph.yaml's placeholder contract addresses from foundry/deployments/localhost.json (falls back to the hardhat deployments path).
Functionality — how & why	How: reads the deployments JSON, rewrites the address: fields in subgraph.yaml in place. Why: was review item #3 from the cross-PR discussion — removes a manual address-editing step before local graph deploy.
Diff vs current develop HEAD	No equivalent on develop.
Recommended merge proposal	Merged as-is — not executed end-to-end (needs a live local deploy artifact) but logic reads correctly.
Actual merge proposal	Soon
Pending proposal	None
Local merge conflict	No
GitHub merge conflict	No

Verification

npm install && npx graph codegen — clean, all data sources + templates + schema typed successfully. npx graph build fails, but the only failure is in src/registry.ts (unrelated file, not touched by this PR — inherited from PR #29, confirmed absent from this PR's own subgraph/-scoped diff). Full detail in the deep-verification comment above.

Local vs. GitHub agree: yes — GitHub's CLEAN/MERGEABLE against feat/din-indexer matches the local analysis; the only nuance GitHub's stat doesn't surface is that this PR's head already carries PR #29's commits, so #29 must land in feat/din-indexer first or its content arrives unreviewed as a side effect of merging #72.