wjx-cli-use
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONPRIVILEGE_ESCALATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill is developed by the official vendor (wjxcom) for the Wenjuanxing platform. All referenced domains (wjx.cn, sojump.cn) and repositories (github.com/wjxcom/*) are official and legitimate.
- [COMMAND_EXECUTION]: The skill operates through a dedicated CLI (
wjx). The provided setup and packaging scripts (setup.sh,pack_skill.sh) use standard system utilities (Node.js, zip, tar, awk) for environment detection and project management. These scripts include robust error handling and path resolution logic. - [EXTERNAL_DOWNLOADS]: The skill guides the user to install the
wjx-clipackage from the public npm registry. It also provides automated setup for Node.js using trusted community scripts from NodeSource and NVM. These downloads are documented for environment preparation and target well-known services. - [REMOTE_CODE_EXECUTION]: The installation guide and
setup.shscript include commands to fetch and execute setup scripts for Node.js (e.g., nvm and nodesource). These are standard practices for developer environment configuration and originate from established sources. - [PRIVILEGE_ESCALATION]: Documentation and scripts include instructions for using
sudoto install system packages and global npm tools. This is typical for managing software on Unix-like systems and is appropriately scoped to installation tasks. - [INDIRECT_PROMPT_INJECTION]: As a tool that processes survey titles, questions, and participant responses, the skill has an inherent attack surface for indirect prompt injection. The skill mitigates this risk by enforcing structured data formats (JSONL), implementing delimiter-based parsing, and providing specific guidance to avoid shell variable expansion traps during data submission.
Audit Metadata