Feedback on your automating-reminders skill #1
Description
I took a look at your automating-reminders skill and wanted to share some thoughts.
Links:
The TL;DR
You're at 100/100, solidly A-grade territory. This is based on Anthropic's best practices for agentic skills. Your strongest area is Spec Compliance (14/15) — your YAML frontmatter is clean and your naming conventions are spot-on. Weakest area is Writing Style (8/10) — mostly minor voice consistency issues that are easy fixes.
What's Working Well
-
Progressive Disclosure Architecture is chef's kiss: You've got a focused SKILL.md (~168 lines) backed by 5 reference files that go exactly one level deep. No nested subdirectories, no token bloat. The "What to Load" section explicitly guides people on progressive loading — that's the kind of thoughtfulness that wins points.
-
Error handling that actually helps: You don't just document the -10024 error; you explain why it happens and give the copy-delete pattern as a workaround. Same with the validation checklist covering permission issues. That's solving real problems.
-
Examples across three languages: JXA, PyXA, PyObjC — you're not assuming one workflow. The recipes for common tasks (creating reminders, moving them) give people templates they can actually use without guessing.
-
Spec compliance is tight: Valid YAML, correct hyphen-case naming, description with good trigger coverage. This is the foundation done right.
The Big One: Voice Consistency
Your writing drifts between imperative (which is right) and second-person occasionally. Example: "ensure Reminders permissions are granted" reads like you're talking to the reader. In progressive disclosure, you're documenting what things are, not giving instructions.
The fix: Scan SKILL.md and rewrite those notes in imperative or passive voice. Change "ensure" → "Reminders permissions must be granted" or "Permissions are required for Reminders access." This bumps Writing Style up ~1 point and makes your skill feel more consistent.
Other Things Worth Fixing
-
Remove "Recommended Modern Approach" from the PyXA header (line 58). That's subjective framing. Just call it PyXA — let the content speak.
-
Trim inline code comments in the quickstart example.
dueDate: new Date(Date.now() + 3*86400*1000), // 3 daysdoesn't need "from now" at the end — you're writing for Claude, not junior devs. -
Add a TOC to your longer reference files if any are creeping toward 100+ lines. Makes grep-ability better and helps with progressive loading signals.
Quick Wins
- Fix voice consistency in notes (+1 point)
- Remove promotional language (+0.5 points)
- Tighten code comments (+1 point)
- These three moves could push you to 103/100 with bonuses
You're already shipping production-ready stuff here. The fixes are genuinely small — mostly editorial polish. Your architecture decisions are solid.
Checkout your skill here: [SkillzWave.ai](https://skillzwave.ai) | [SpillWave](https://spillwave.com) 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.