Security : NodeStaking DID squatting + nodeIds growth + slashing gap

[philpof102-svg]

Three combined NodeStaking findings :

1. DID squatting (HIGH) — registerNode accepts any bytes32 hash, no DID controller proof. ~$90 in $GITLAWB squats 100 future DIDs.
2. nodeIds memory leak (MEDIUM, amplifies Security HIGH: NodeStaking.depositRevenue O(n) loops gas DoS #7) — unstake never removes the entry. After 5 years of churn → guaranteed gas-out even at low active count.
3. Slashing scope gap (INFO) — docs promise 10-100% slashing but no slash() function exists in code.

Full PoC + fix patterns in BUG_REPORT_nodestaking_did_squatting.md.

5th PR in continuous Gitlawb audit. Helping ship Phase 8 mainnet safe. Reporter @philpof102-svg.

[philpof102-svg]

Upgraded with merge-ready fix (src/GitlawbNodeStaking.sol, +53 lines).

Two fixes bundled :

1) DID squatting (opt-in binding)

interface IDIDRegistryView {
    function didToOwner(bytes32) external view returns (address);
}

IDIDRegistryView public didRegistry;

function registerNode(...) external {
    ...
    if (address(didRegistry) != address(0)) {
        if (didRegistry.didToOwner(nodeDidHash) != msg.sender) revert NotDIDOwner();
    }
    ...
}

function setDIDRegistry(address _r) external { ... } // onlyOwner

Zero-address default preserves legacy behavior — existing deployments don't break, and you opt in via setDIDRegistry(GitlawbDIDRegistry).

2) nodeIds swap-and-pop (memory leak → O(1) cleanup)

mapping(bytes32 => uint256) private nodeIdIndex; // 1-based

// registerNode
nodeIds.push(nodeDidHash);
nodeIdIndex[nodeDidHash] = nodeIds.length;

// unstake
uint256 idx1 = nodeIdIndex[nodeDidHash];
if (idx1 != 0) {
    uint256 last = nodeIds.length - 1;
    uint256 idx = idx1 - 1;
    if (idx != last) {
        bytes32 moved = nodeIds[last];
        nodeIds[idx] = moved;
        nodeIdIndex[moved] = idx + 1;
    }
    nodeIds.pop();
    nodeIdIndex[nodeDidHash] = 0;
}

Pre-fix nodes have nodeIdIndex == 0 so the cleanup is a no-op for them — backward compatible.

Combined with the PR #7 loop-reduction in depositRevenue, this gets the protocol to a place where revenue distribution stays cheap even after long operator churn.