declarative-agent-developer
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFECOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [Command Execution]: The skill uses the wiqd CLI for all agent lifecycle tasks, including creation, validation, and deployment. It also utilizes curl and powershell commands for environment setup and discovery.
- [Remote Code Execution]: The skill employs npx to run the mcp-remote-client utility for MCP server authentication, which involves executing a remote package as part of the authentication flow.
- [Dynamic Execution]: During tool discovery, the skill uses Python heredoc scripts to parse JSON responses. It also performs file system operations like moving and removing directories to scaffold project structures.
- [Indirect Prompt Injection Surface]: The skill processes external data such as OpenAPI specifications and MCP tool definitions. While this presents an attack surface, the skill mitigates it through mandatory validation with wiqd agent validate and a strict Detect-Inform-Ask protocol to ensure user review.
Audit Metadata