react

I took a look at your react skill and wanted to share some thoughts.

Links:

• GitHub
• Your agentic skill in the SkillzWave marketplace

The TL;DR

You're at 81/100, solidly in B territory—that's good work. This evaluation is based on Anthropic's skill best practices rubric. Your strongest area is Ease of Use (22/25) with excellent trigger coverage and clear workflows. The weaker spots are Progressive Disclosure Architecture (23/30) and Utility (17/20), both dragged down by some redundancy and missing error-handling guidance.

What's Working Well

• Trigger coverage is chef's kiss. You've got both intent triggers ("React hooks", "component patterns") and artifact triggers (regex for .tsx files). This skill activates when it should.
• Clear routing table maps user intents to specific actions—that's practical and saves users from figuring out what to do next.
• Library IDs table is smart. Gives the skill the Context7 library IDs upfront so it doesn't waste tokens resolving them.
• Success criteria checklist in the quick_start gives users a way to validate they got what they needed.

The Big One: Redundancy in PDA

Your anti_patterns section duplicates your constraints section—both cover "don't use useEffect for derived state", "no 150+ line components", etc. This wastes tokens and tanks your PDA score.

Fix: Delete the entire <anti_patterns> section. Your constraints already say it all. That's a quick +2 points on token economy (takes you from 7/10 to 9/10 in that category).

Similarly, your <component_pattern> section in SKILL.md mirrors the template file almost exactly. Reference the template instead: "See templates/component.tsx.md for scaffolding." This cuts ~20 lines of fat.

Other Things Worth Fixing

1. Second-person voice in quick_start — You write "Fetch docs for your task" but spec prefers imperative. Change to "Fetch docs for the current task." Small thing, +0.5 points on writing style.

2. Missing error handling guidance — No fallback if Context7 is unavailable. Add a note like: "If Context7 unavailable, check node_modules types or reference the research section patterns." This bumps feedback_loops from 3/4 to 4/4.

3. Description needs more trigger phrases — You've got 1-2 in the description; add 3-4 more short phrases to help discoverability. Something like "React components", "hooks patterns", "TypeScript React".

4. Add a TOC for files over 100 lines — At ~274 lines, a simple table of contents at the top would make navigation easier and score better on navigation signals.

Quick Wins

• Delete <anti_patterns> section entirely (+2 points)
• Replace <component_pattern> with template reference (+1 point)
• Fix second-person voice in quick_start (+0.5 points)
• Add error-handling fallback guidance (+1 point)
• Expand description triggers and add TOC

These fixes should get you into the 85-87 range without major restructuring.

Checkout your skill here: SkillzWave.ai | SpillWave We have an agentic skill installer that install skills in 14+ coding agent platforms. Check out this guide on how to improve your agentic skills.