Escrow `fund()` allows issue-id griefing via front-running before the legitimate sponsor funds

[chonilius]

Background / Context

contracts/escrow/src/lib.rs::fund() uses issue_id (a u64 chosen off-chain by mergefi-backend/the sponsor's wallet) as the sole key into persistent storage (DataKey::Escrow(issue_id)), and rejects a second fund call on the same id with AlreadyFunded. Critically, fund performs no admin check and no validation that the caller is the "intended" sponsor for that issue — any address can call fund(issue_id, sponsor=attacker, token=<attacker-chosen>, amount=1, deadline=<attacker-chosen>) for any issue_id, as long as they supply a valid require_auth signature for themselves as sponsor.

Problem Statement

Because Stellar issue ids are presumably predictable (sequential GitHub issue numbers, or a deterministic hash the backend computes ahead of time), an attacker who front-runs the real sponsor's fund transaction — or simply calls fund proactively for issue ids they anticipate being funded — permanently claims that issue_id slot with a trivial deposit (amount=1) of a token of their choosing. The legitimate sponsor's later fund call then reverts with AlreadyFunded, and there is no top-up or re-fund-different-id path described in the README. This is a low-cost, high-impact griefing vector against a system whose entire value proposition is "sponsors fund GitHub issues."

Requirements

• Analyze the full griefing surface: is issue_id predictable/guessable before the real fund call lands? Model the mempool/submission-ordering assumptions on Stellar/Soroban (there is no public mempool in the traditional sense, but front-running via fee/priority is still analyzed differently — investigate Soroban's actual transaction ordering guarantees).
• Propose and implement a mitigation: options include (a) binding fund to a specific expected sponsor address set at issue-creation time by the admin/oracle, (b) a commit-reveal or admin-pre-registration step before fund is callable, (c) an admin-callable reassign/force_refund_and_reopen recovery path for squatted ids, or another design you can justify as superior.
• Add regression tests in contracts/escrow/src/test.rs that reproduce the griefing scenario and prove the fix closes it (attacker funds first, legitimate sponsor's transaction should still succeed, or attacker's low-value funding should be recoverable without requiring off-chain intervention).
• Document the chosen trust model change in the README's escrow section.

Acceptance Criteria

• Written analysis of whether/why issue_id squatting is exploitable given Soroban's actual ordering model (not just Ethereum-style mempool assumptions)
• A concrete, implemented mitigation in contracts/escrow/src/lib.rs
• New tests proving the attack fails post-fix and existing legitimate flows (README's funding flow) still pass
• cargo test --workspace green, make build still produces valid wasm
• README updated to reflect the new trust assumptions for fund

Technical Notes / Hints

• Relevant code: contracts/escrow/src/lib.rs lines ~54-88 (fund), types.rs DataKey::Escrow(u64).
• Consider whether the fix should also apply symmetrically to create_milestone (contracts/milestones/src/lib.rs) and deposit's implicit pool-creation-on-first-deposit (contracts/maintenance-pool/src/lib.rs) — file separate issues if the fix diverges per contract, but note the shared pattern here.

Difficulty Justification

This requires understanding Soroban/Stellar's actual transaction-ordering and mempool semantics (not assumed from EVM intuition), reasoning about an economic griefing attack with no simple "add a require" fix, and designing a new access-control primitive that doesn't break the permissionless, oracle-driven trust model the rest of the contract relies on. Getting the fix wrong (e.g., adding an admin-only fund) would break the sponsor-initiated funding flow the whole product depends on.

[rupesh-kumar-sah]

Hello @maintainers,

I would love to take this on. My planned approach:

• Implement Escrow fund() allows issue-id griefing via front-running before the legitimate sponsor funds with explicit require_auth checks, secure state storage, and gas optimization.
• Cover contract calls with hermetic unit tests verifying events and token balances.

Looking forward to contributing—please assign whenever convenient!

[Smithis]

@Smithis has applied to work on this issue as part of the Stellar Wave Program's 9th wave.

fund() issue-id griefing via front-running — same guard/reliability class as shipped Soroban escrow fixes. Same-day turnaround with regression test.

ℹ️ Repo Maintainers: To accept this application, review their application or assign @Smithis to this issue.

[Leumas887]

@Leumas887 has applied to work on this issue as part of the Stellar Wave Program's 9th wave.

After carefully reviewing the issue, I have a clear understanding of the requirements and believe I can contribute effectively to the implementation. The task is well aligned with my current skills, and I’m prepared to put in the necessary effort to deliver a reliable solution that meets the project’s expectations. I would appreciate the opportunity to take responsibility for this issue. Thank you for your consideration.

ℹ️ Repo Maintainers: To accept this application, review their application or assign @Leumas887 to this issue.

[Wilfred007]

@Wilfred007 has applied to work on this issue as part of the Stellar Wave Program's 9th wave.

Can I work on this task?

ℹ️ Repo Maintainers: To accept this application, review their application or assign @Wilfred007 to this issue.

[drips-wave]

Congratulations, @Wilfred007! 🎉 Your application was accepted by the repo's maintainers, and the issue is due on September 30, 2026.

🧑‍💻 @Wilfred007: Please resolve the issue such that the repo's maintainers have enough time to review your contribution before the due date. You'll earn Points for completing the issue on-time, which will make you eligible for a share of the Stellar Wave Program's reward pool.

Warning

When opening a PR, please link it to this issue to ensure it gets tracked accurately. Points are awarded when this issue is marked as completed by the maintainer.

🤠 Repo maintainers: Please keep an eye on the contributor's progress and review their work before the due date. You can manage this issue, including adjusting its complexity and points, here.

🌊 Happy Wave 🌊