The Agent Skills Directory

[SAFE]: Structured Critique Methodology. The skill implements a robust framework for evaluating user interfaces using established UX dimensions such as visual hierarchy, emotional resonance, and cognitive load. It relies on provided reference documents to ensure consistency in scoring and evaluation.
[SAFE]: Secure Context Integration. The skill reads metadata from .github/copilot-instructions.md to customize analysis based on the project's brand and audience. This data ingestion is handled within the local project context for generation purposes and does not involve risky tool execution or data exfiltration.
Ingestion points: .github/copilot-instructions.md (Design Context section).
Boundary markers: The skill instructs the agent to search specifically for the Design Context section, minimizing exposure to other instructions.
Capability inventory: The skill recommends design-related commands (e.g., /polish, /audit, /harden) to the user; it does not contain code for autonomous shell execution, network calls, or file system modifications.
Sanitization: Not applicable as the skill performs heuristic analysis without executing ingested content as code.

critique