Pre-auditThree days of senior audit time for $1,500, credited in full toward your Zealynx audit.3 audit days, credited toward your auditSee how →
All case studies

Direct Zealynx engagement

Dripster contracted Zealynx directly. Zealynx acted as Smart contract security audit team.

Dripster · Prediction Markets · Apr 2026

Dripster: Verifying Storage Before a UUPS Upgrade

A UUPS vault review found misleading storage-slot annotations. The fix corrected the map and added automated layout validation to CI.

Invariant under test

A UUPS upgrade must preserve every existing proxy storage slot, regardless of comments or inherited storage assumptions.

The answer in 30 seconds

Dripster directly engaged Zealynx to audit a seven-file Solidity vault that coordinated leveraged Polymarket positions through a UUPS-upgradeable proxy. The reviewed contract used OpenZeppelin v5, but all 28 inline storage-slot annotations described the older sequential-storage model. The implementation was correct in V1; the danger was that a later developer could trust the comments, place a V2 variable over an existing admin slot, and corrupt proxy governance during an upgrade.[1][2]

Zealynx checked the proxy's actual slots with Foundry, showed that globalAdmin occupied slot 0 rather than the documented slot 4, and recommended replacing assumption-based planning with generated layout validation. The report records the Medium issue as fixed: the annotations were corrected and the OpenZeppelin Foundry Upgrades plugin now validates layout in CI.[1]

This is a direct Zealynx technical engagement narrative based on public evidence, not a customer transformation story. It does not claim exploitation, deployment of a V2 upgrade, monetary exposure, customer sentiment, or continuing assurance.

The buyer decision: is the proxy safe to upgrade?

Dripster's reviewed vault held position state across a fourteen-state lifecycle and used separate admin roles for global control, emergency response, and application operations. Its UUPS design allowed implementation changes while the proxy retained state.[1][2]

For a founder or smart contract lead, the relevant decision is not simply whether the current implementation passes tests. It is whether the team can add V2 state without overwriting the proxy's existing authority, accounting, or position data.

The invariant was:

A UUPS upgrade must preserve every existing proxy storage slot, regardless of comments or assumptions about inherited contracts.

The dangerous assumption: inherited contracts still consume slots 0–3

The source comments assumed that UUPSUpgradeable, ReentrancyGuardUpgradeable, PausableUpgradeable, and EIP712Upgradeable occupied four sequential slots before Dripster's variables. That model matched an older OpenZeppelin layout, but OpenZeppelin v5 moved inherited state into ERC-7201 hash-derived namespaces.[1]

Dripster's own variables therefore began at slot 0, not slot 4. The report found every one of the 28 slot annotations shifted by four positions. The storage gap was also documented as slots 32–65 even though it actually covered slots 24–57.[1]

Nothing in V1 read those comments at runtime. The failure would emerge later, when a developer treated documentation as the storage map for an upgrade.[1]

How Zealynx tested the storage-preservation invariant

1. Separate runtime correctness from upgrade safety

The review first established that V1's actual layout and storage gap were valid. This prevented a documentation defect from being exaggerated into a present runtime collision.[1]

2. Inspect the proxy's slots directly

A Foundry proof used vm.load against the proxy and verified the four admin addresses in slots 0–3. In particular, the test read slot 0 and matched it to globalAdmin, disproving the slot-4 annotation.[1]

3. Model a plausible V2 planning error

The report followed the misleading comment into an upgrade scenario: a developer assumes slots 0–3 are reserved by inherited contracts, places a new V2 variable at slot 0, and calls upgradeToAndCall. The new write then overlaps globalAdmin in proxy storage.[1]

4. Replace comments as the control boundary

Correcting the annotations addressed the immediate defect, but comments can drift again. The recommended durable control was to run forge inspect LeveragedPredictionVaultV1 storage-layout as part of upgrade verification. The recorded fix goes further by adding OpenZeppelin Foundry Upgrades layout validation to CI.[1]

The pivotal mechanism: valid code plus a false map

The issue required no vulnerability in the UUPS proxy itself. It came from a mismatch between three representations:

actual proxy storage
        ≠
source-code slot comments
        ≠
a future developer's V2 allocation plan

A developer following the comments could interpret slot 0 as unused by Dripster. In reality, slot 0 already held globalAdmin. Writing a new variable there could change or disable the protocol's highest authority after an upgrade.[1]

This is why storage documentation is security-sensitive even when the compiler ignores it: upgradeable systems carry today's layout into tomorrow's implementation.

Before, intervention, and verified after-state

Before the reviewZealynx interventionVerified after-state in the report
All 28 slot annotations were shifted by four positionsCompared the OpenZeppelin v5 namespaced model with Dripster's sequential commentsSlot annotations were corrected
globalAdmin was documented at slot 4 but stored at slot 0Read the proxy slots with a Foundry proofThe corrected map reflects the actual layout
The storage gap was documented as slots 32–65 instead of 24–57Traced the error into V2 capacity planningLayout validation was added to CI using the OpenZeppelin Foundry Upgrades plugin
A future upgrade could rely on stale commentsRecommended generated storage-layout inspection before deploymentThe finding is recorded as fixed

“Fixed” means the April 2026 report records corrected annotations and CI layout validation. The public record does not independently show a CI run, a production deployment, a completed V2 upgrade, or the safety of later code.[1]

A reusable upgrade review model

Before approving an upgradeable-contract release:

  1. Generate the layout: derive storage positions from compiler or upgrade-tool output rather than comments.
  2. Compare versions: check the current deployed layout against the proposed implementation, including inherited and namespaced storage.
  3. Read critical slots: verify admin, implementation, accounting, and asset-reference slots against the proxy where feasible.
  4. Test initialization: exercise upgradeToAndCall with the real migration path and assert old state remains unchanged.
  5. Validate gaps: confirm gap ranges from generated output and update them only through a reviewed allocation plan.
  6. Pin dependencies: treat an OpenZeppelin major-version change as a storage-model review trigger.
  7. Automate the gate: make incompatible layouts fail CI rather than depend on reviewer memory.
  8. Keep comments secondary: document the layout for humans, but never make comments the source of truth.

The transferable lesson is that upgrade safety is a comparison between concrete layouts. Inheritance assumptions and hand-maintained slot numbers are useful explanations, not sufficient release controls.

Impact and evidence limits

The strongest supported outcome is control-specific: the report records correction of the misleading storage map and automated compatibility validation in CI.[1]

The issue was rated Medium because a future V2 developer relying on the comments could overwrite one of four admin addresses or misjudge storage-gap capacity. The report explicitly states that V1 had no runtime impact and that no contract logic depended on the annotations.[1]

The wider review reported 17 findings: two Medium, seven Low, and eight Informational. Nine were fixed and eight acknowledged.[1][2] Those totals do not establish that the complete system, a deployed proxy, or later versions are secure. No public source establishes exploitation, a storage collision in production, customer loss, avoided loss, or a scoped monetary value.

Evidence and attribution

  • Canonical technical record: the public Zealynx report identifies Dripster, Zealynx's audit role, the reviewed scope and commit, the 28 incorrect annotations, the Foundry proof, the upgrade scenario, and the recorded fix.[1]
  • Public audit record: Zealynx's engagement page reports the seven-file scope, direct audit team, finding distribution, and aggregate remediation state.[2]
  • Attribution class: direct Zealynx engagement. The report identifies Zealynx as the audit provider and no public evidence identifies a delivery partner.[1]
  • Evidence cutoff: the public report dated April 28, 2026 and the corresponding audit record.
  • Limitations: provider-authored evidence supports technical scope and recorded remediation, but not deployment, customer sentiment, business impact, exploitation, or continuing assurance.

Planning a UUPS or proxy upgrade?

Zealynx can compare deployed and proposed layouts, test migration calls, inspect privileged slots, and turn storage compatibility into a release gate. Explore smart contract security audits, review the security methodology, or request an upgrade-focused assessment.

Frequently asked questions

Did OpenZeppelin v5 corrupt Dripster's storage?

No. The report says V1's actual layout and gap were correct. The problem was that comments described the wrong model, which could mislead a future upgrade.[1]

Why did the first Dripster variable start at slot 0?

OpenZeppelin v5 stores the inherited upgradeable modules' state in ERC-7201 hash-derived namespaces rather than consuming the sequential slots assumed by the comments. Dripster's sequential variables therefore began at slot 0.[1]

How was the actual layout verified?

The report includes Foundry tests that use vm.load to read proxy slots directly and match slots 0–3 to the four admin addresses.[1]

What could happen if V2 reused slot 0?

Slot 0 held globalAdmin. A V2 variable written there through the proxy could overwrite that authority, potentially changing control or disabling administrative access.[1]

How was the issue fixed?

The report states that the slot annotations were corrected and the OpenZeppelin Foundry Upgrades plugin now validates the storage layout in CI.[1]

Does the fix prove a future upgrade is safe?

No. Layout validation addresses the documented compatibility hazard. Each proposed implementation, migration call, dependency change, and deployment still requires review.

Sources

[1] https://raw.githubusercontent.com/ZealynxSecurity/audits/main/web3/2026-04-zealynx-dripster.pdf — Dripster Smart Contract Audit [2] https://zealynx.io/audits/dripster/leveraged-prediction-vault-q2-2026 — Dripster Leveraged Prediction Vault audit record