Skip to content

Security: DevQwinB/soroban-keeper-network

Security

SECURITY.md

Security Policy

Supported Versions

Version Supported
main (latest) Yes
develop No (pre-release)
All others No

Reporting a Vulnerability

Please do NOT report security vulnerabilities via GitHub Issues, Pull Requests, or Discussions. Public disclosure before a fix is available puts all users at risk.

Responsible Disclosure Process

  1. Email your report to: security@soroban-keeper.network
    (Until the domain is registered, DM @Andreschuks101 on GitHub with subject line: [SECURITY] Keeper Network)

  2. Include in your report:

    • A description of the vulnerability
    • Steps to reproduce
    • Potential impact (funds at risk? DoS? privilege escalation?)
    • Your suggested severity (critical / high / medium / low)
    • Any proof-of-concept code (kept confidential)
    • Whether you want to be credited in the fix announcement
  3. We will respond within 24 hours acknowledging receipt.

  4. Timeline:

    • Critical (funds at direct risk): patch within 7 days
    • High: patch within 14 days
    • Medium/Low: patch within 30 days
  5. Disclosure: we will coordinate a public disclosure date with you. We request a minimum 90-day embargo for critical issues to allow users to upgrade.

  6. Credit: security researchers who report valid vulnerabilities will be credited in the release notes (unless they prefer anonymity).

Scope

In Scope

  • contracts/keeper-registry/ — all on-chain logic
  • Reward escrow and distribution logic
  • Authentication and authorization bypass
  • Reentrancy attacks
  • Storage collision or poisoning
  • Integer overflow/underflow that leads to fund loss
  • Upgrade mechanism abuse
  • Event spoofing or manipulation

Out of Scope

  • Issues in dependencies that are already publicly disclosed (report upstream)
  • Theoretical attacks with no practical exploit
  • Spam / griefing that only affects the attacker economically
  • The off-chain keeper bot example (examples/keeper-bot/) — this is illustrative code, not production infrastructure

Smart Contract Security Model

The KeeperRegistry contract follows these security principles:

  • Checks-Effects-Interactions (CEI) — all state mutations happen before token transfers
  • Auth gating — every write function requires address.require_auth()
  • No re-entrancy via Soroban's synchronous execution model — Soroban does not support cross-contract callbacks that could re-enter; however, we follow CEI regardless
  • Bounded iteration — no loops over unbounded collections in storage
  • Overflow protectionoverflow-checks = true in Cargo profile + checked_* arithmetic in hot paths
  • Principle of least privilege — keepers can only execute tasks they have claimed; admin cannot steal escrowed task rewards

Known Limitations (by Design)

  1. No on-chain execution verification (MVP) — The registry credits rewards to keepers who submit any proof bytes. A dishonest keeper could claim-and-submit-fake-proof. Mitigation: Phase 2 adds an optional on-chain verifier callback that target protocols can implement.

  2. Admin trust — The admin can pause the contract and upgrade the WASM. This is intentional for the MVP. Phase 2 will replace admin with a governance contract requiring token-holder votes.

  3. Oracle-less deadline enforcement — Deadlines are compared against env.ledger().timestamp() which is set by the Stellar consensus network. This is a trust assumption on the Stellar validator set (acceptable for this use case).

Bug Bounty

A formal bug bounty program will be announced after the first production deployment and audit. Until then, we offer:

  • Critical bugs (direct fund loss): up to $5,000 USD equivalent in XLM (at maintainer discretion, subject to project treasury availability)
  • High bugs: public credit + potential discretionary reward
  • Medium/Low bugs: public credit in release notes

Audit Status

Date Auditor Scope Report
Planned Q4 2026 TBD keeper-registry v1.0 TBD

Pre-audit internal reviews are tracked in GitHub Issues with the security label.

There aren't any published security advisories