skills/wjxcom/wjx-ai-kit/wjx-mcp-use/Gen Agent Trust Hub

wjx-mcp-use

Pass

Audited by Gen Agent Trust Hub on Nov 2, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill provides a robust set of tools for survey creation, response management, and data analysis specifically for the WJX platform. All analyzed functionalities align with the stated purpose of the skill.
  • [COMMAND_EXECUTION]: The skill instructions in references/host-routing.md suggest that the agent should perform an environment check by running wjx --version. This is a standard practice for verifying the installation of necessary CLI tools and does not pose a security risk.
  • [REMOTE_CODE_EXECUTION]: The skill mentions the installation of the vendor's own CLI tool via npm install -g wjx-cli@latest. This is presented as a user-initiated setup step and targets the official package from the skill author, falling within standard operational procedures for developer-focused skills.
  • [CREDENTIALS_UNSAFE]: The documentation references WJX_API_KEY and MCP_AUTH_TOKEN. It explicitly guides users to set these as environment variables in the host configuration and provides official platform URLs for key retrieval, demonstrating a proactive approach to secret management rather than hardcoding credentials.
  • [DATA_EXFILTRATION]: While the skill includes tools for data management (e.g., query_responses, upload_file), these are restricted to the context of the user's authenticated platform account. The skill also requires explicit user confirmation for high-risk operations, such as consuming realtime response queues, which prevents unauthorized data manipulation.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes external data (survey responses). However, it uses structured data formats (JSONL, XML DSL) and specifies validation steps for these inputs. The agent is instructed to follow a strict protocol (Found intent -> Pre-check -> Plan -> Confirm -> Execute), which mitigates the risk of untrusted data influencing agent behavior without oversight.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 2, 2026, 12:59 AM