gate-dex-wallet-cli

Pass

Audited by Gen Agent Trust Hub on Sep 19, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPRIVILEGE_ESCALATIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill downloads an executable binary from a remote server to enable its core functionality. The setup.sh script and the SKILL.md frontmatter both fetch the gate-dex binary from https://gate-dex-cli.gateweb3.cc.
  • [REMOTE_CODE_EXECUTION]: The skill downloads and subsequently executes an external binary file. This gate-dex tool is the primary mechanism through which all wallet and transaction operations are performed.
  • [COMMAND_EXECUTION]: The agent is instructed to execute various shell commands through the gate-dex CLI, including high-risk operations like login, send, sign-msg, and sign-tx.
  • [PRIVILEGE_ESCALATION]: The setup.sh script applies executable permissions (chmod +x) to the gate-dex binary file after it is downloaded.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests external data from the blockchain that could potentially contain malicious instructions intended to influence the agent's behavior.
  • Ingestion points: The agent reads token balances, transaction history, and swap order details via the gate-dex tokens and gate-dex tx-history commands (documented in references/asset-query.md).
  • Boundary markers: There are no explicit delimiters or specific instructions provided to the agent to treat blockchain-sourced data as untrusted or to ignore embedded instructions.
  • Capability inventory: The skill possesses significant blockchain write capabilities, including the ability to transfer funds (send), sign arbitrary hex messages (sign-msg), and sign raw transactions (sign-tx).
  • Sanitization: There is no evidence of sanitization or escaping of data retrieved from the blockchain before it is presented to the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 19, 2026, 09:04 AM