create-architecture-doc
Pass
Audited by Gen Agent Trust Hub on Nov 10, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill implements a structured documentation workflow using local templates and resources. It focuses on drafting architectural plans with mandatory user feedback, which ensures human oversight. All external references (such as Vercel, AWS, and standard frameworks) are used as examples of well-known technologies and established services.\n- [INDIRECT_PROMPT_INJECTION]: The skill includes capabilities to analyze existing codebases and documentation, particularly in the 'brownfield' and 'fullstack' templates. This creates an indirect prompt injection surface if the analyzed external project files contain malicious instructions.\n
- Ingestion points: Analyzing user-provided project files and repositories as described in the 'Starter Template or Existing Project' section of
architecture-tmpl.yamland the 'Existing Project Analysis' section ofbrownfield-architecture-tmpl.yaml.\n - Boundary markers: No specific delimiters or instructional blocks are used to isolate potentially untrusted project data from the agent's internal logic.\n
- Capability inventory: The agent reads and processes the contents of project directories to generate summaries, data models, and architectural schemas.\n
- Sanitization: No specific sanitization or filtering logic is defined for the content of the analyzed files.
Audit Metadata