zksync

Fail

Audited by Gen Agent Trust Hub on Aug 4, 2026

Risk Level: HIGHREMOTE_CODE_EXECUTIONCREDENTIALS_UNSAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The documentation and skill body instruct users to execute remote scripts by piping them directly to the bash interpreter. \n
  • Evidence: curl -L https://raw.githubusercontent.com/matter-labs/foundry-zksync/main/install-foundry-zksync | bash in SKILL.md and examples/deploy-contract/README.md.\n
  • Risk: This practice allows for unverified code execution from an external server, bypasses standard auditing, and represents a significant supply chain vulnerability.\n- [CREDENTIALS_UNSAFE]: The skill relies on the PRIVATE_KEY environment variable for critical blockchain operations across all examples and templates.\n
  • Evidence: Usage of process.env.PRIVATE_KEY in templates/zksync-deploy.ts and examples/bridge/README.md.\n
  • Risk: Exposure of the environment variables or the .env file would compromise the associated blockchain accounts, providing an attacker with full control over the funds.\n- [COMMAND_EXECUTION]: The skill integrates multiple shell-based toolchains for compiling and deploying smart contracts.\n
  • Evidence: Use of npx hardhat deploy-zksync and forge commands.\n
  • Risk: The agent has the capability to trigger system-level commands, which could be abused if the agent receives malicious instructions.\n- [EXTERNAL_DOWNLOADS]: The skill fetches various dependencies and references external RPC endpoints for blockchain interactions.\n
  • Evidence: Installation of zksync-ethers and hardhat via npm, and connections to https://mainnet.era.zksync.io.\n- [PROMPT_INJECTION]: The templates provided by the skill have an attack surface for indirect prompt injection via unsanitized data inputs used in blockchain transactions.\n
  • Ingestion points: constructorArgs in templates/zksync-deploy.ts and args in sendWithPaymaster function.\n
  • Boundary markers: None identified in the provided templates.\n
  • Capability inventory: High-privilege actions including contract deployment and transaction execution (wallet.sendTransaction).\n
  • Sanitization: No input validation or escaping is performed before data is passed to blockchain interaction functions.
Recommendations
  • HIGH: Downloads and executes remote code from: https://raw.githubusercontent.com/matter-labs/foundry-zksync/main/install-foundry-zksync - DO NOT USE without thorough review
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Aug 4, 2026, 12:32 AM
Security Audit — agent-trust-hub — zksync