technology-stack-blueprint-generator
Pass
Audited by Gen Agent Trust Hub on Feb 25, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No security vulnerabilities or malicious patterns were identified. The skill performs standard codebase analysis using localized configuration file scanning and pattern detection.
- [PROMPT_INJECTION]: The skill processes codebase content, which is an inherent surface for indirect prompt injection. However, this is consistent with its primary documentation purpose. Ingestion points: Scans dependencies, project files, and source code content across the repository (SKILL.md). Boundary markers: The generated prompt organizes the analysis logically but does not use specific delimiters to sandbox ingested code. Capability inventory: The skill is limited to analysis and saving results as a blueprint file. Sanitization: The skill relies on the underlying agent's standard safety protocols when interpreting code files.
Audit Metadata