workflow-skill-architect
Pass
Audited by Gen Agent Trust Hub on Mar 28, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill serves as an architectural assistant for generating markdown-based instructions for AI agents. It does not perform network operations (other than displaying documentation links), access sensitive local files, or execute shell commands.
- [EXTERNAL_DOWNLOADS]: The skill includes several references to official developer documentation from Anthropic (Claude) and Cursor. These are well-known, trusted domains (claude.com, cursor.com) used for providing legitimate guidance on skill authoring.
- [DATA_EXPOSURE]: No hardcoded credentials, API keys, or sensitive file paths were detected. The skill operates on user-supplied workflow descriptions and does not attempt to harvest environment data.
- [PROMPT_INJECTION]: The instructions are task-oriented and do not contain attempts to bypass AI safety filters, extract system prompts, or override agent constraints. The "Default to subagent" and "orchestrator context window" instructions are legitimate design patterns for optimizing LLM performance.
- [INDIRECT_PROMPT_INJECTION]: While the skill processes user-supplied workflow steps to generate content, this is the intended primary purpose of the tool. It uses structured XML templates to delineate user input, which is a recommended practice for reducing the risk of accidental instruction confusion.
Audit Metadata