gate-exchange-earn
Pass
Audited by Gen Agent Trust Hub on Apr 30, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The skill employs specific directives to ensure the agent follows mandatory runtime rules and safety guardrails, including a strict 'Action Draft' protocol that requires the user to explicitly confirm any write operations in the turn immediately preceding execution. It also possesses a surface for indirect prompt injection as it processes data from external API responses. 1. Ingestion points: Product lists and market rates fetched via 'gate-cli' (SKILL.md). 2. Boundary markers: The 'Action Draft' serves as a logical confirmation boundary. 3. Capability inventory: Subprocess calls to 'gate-cli' for financial actions. 4. Sanitization: Not explicitly implemented for external API content.
- [DATA_EXFILTRATION]: Account and position data are accessed exclusively via the official Gate.io CLI. The skill correctly manages credentials using environment variables and provides clear warnings against sharing API secrets in the conversation.
- [EXTERNAL_DOWNLOADS]: The skill facilitates the download of the 'gate-cli' binary from official vendor repositories on GitHub. The 'setup.sh' script implements checksum verification to ensure the integrity of the downloaded file.
- [COMMAND_EXECUTION]: The skill invokes 'gate-cli' for account data retrieval and financial transactions. All execution-oriented commands are protected by a mandatory confirmation gate. The installation script may use 'sudo' for tool installation if local directory access is restricted.
- [REMOTE_CODE_EXECUTION]: The skill downloads and executes a binary from a remote source. This source is the official repository of the vendor, and the process is governed by checksum verification and explicit usage documentation.
Audit Metadata