gmgn-wallet-review
Pass
Audited by Gen Agent Trust Hub on Nov 14, 2026
Risk Level: SAFECREDENTIALS_UNSAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [CREDENTIALS_UNSAFE]: The instructions direct the agent to request a sensitive key from the user and then execute
gmgn-cli config --apply <KEY>. This practice can lead to the exposure of secrets in chat logs, shell history, or process monitoring tools. - [COMMAND_EXECUTION]: The skill's primary functionality is built on executing shell commands, including the
gmgn-cliutility and data processing tools likejqandpython3. While the skill includes regex validation for arguments, arbitrary command execution remains a potential risk factor. - [EXTERNAL_DOWNLOADS]: The skill references and recommends the installation of the
gmgn-clipackage via the npm registry. This is a vendor-specific tool required for the skill's operation. - [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from blockchain transactions and token metadata. It acknowledges this risk by instructing the agent to treat third-party text as data and avoid following any instructions embedded within it. The ingestion occurs through CLI command outputs which are then processed by the agent.
- [DYNAMIC_EXECUTION]: The skill dynamically generates HTML and JavaScript for rendering wallet reports. It also uses
python3 -cto execute dynamically constructed scripts for data processing. The skill includes XSS mitigation instructions, such as escaping characters, for the HTML generation process.
Audit Metadata