Content
50%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This skill provides a well-structured critique framework with a useful output template, but it over-explains feedback principles Claude already understands and lacks concrete worked examples showing the framework applied to an actual design. The workflow could benefit from explicit sequencing and a validation step to confirm context before delivering the critique.
Suggestions
Add a concrete worked example showing the critique framework applied to a hypothetical design (e.g., a login screen), demonstrating the expected output quality and specificity level.
Remove or significantly trim the 'How to Give Feedback' section — Claude already knows these general feedback principles; instead, encode them as brief inline reminders within the output template.
Add an explicit workflow sequence: 1) Identify input type, 2) Confirm context/stage with user if missing, 3) Apply framework, 4) Generate output — with a checkpoint to verify you have sufficient context before proceeding.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill is moderately efficient but includes some unnecessary padding. The 'How to Give Feedback' section explains general feedback principles Claude already knows (be specific, explain why, suggest alternatives). The 'Tips' section at the end largely repeats information from 'What I Need From You'. The critique framework dimensions are useful but could be more concise. | 2 / 3 |
Actionability | The skill provides a clear output template and structured critique framework, which is good. However, it lacks concrete executable examples — there's no example of an actual critique applied to a real or hypothetical design. The framework is more of a checklist than executable guidance, and the Figma integration instructions are vague ('pull the design from Figma' without specifying how). | 2 / 3 |
Workflow Clarity | The workflow is implicit rather than explicit: receive design → gather context → apply framework → produce output. The steps for handling different input types (Figma URL vs file vs description) are mentioned but not clearly sequenced. There's no validation checkpoint — e.g., confirming context/stage with the user before proceeding with the critique, or verifying the output covers all requested focus areas. | 2 / 3 |
Progressive Disclosure | The skill references CONNECTORS.md appropriately, but the content is somewhat monolithic — the full critique framework, output template, feedback guidelines, and tips are all inline. The output template alone is ~25 lines that could be in a separate reference file. For a skill of this length (~100 lines), the organization is acceptable but not optimal. | 2 / 3 |
Total | 8 / 12 Passed |