delegation-prompt-crafter

I've been reading through how you structured the prompt delegation flow—the way you layer context and scaffolding to guide users through complex decisions is genuinely clever. At 99/100, that one missing point is probably telling you something worth fixing.

Links:

• GitHub
• Your agentic skill in the SkillzWave marketplace

The TL;DR

You're at 99/100, solid A grade. This is based on Anthropic's skill best practices. Your strongest areas are Writing Style (10/10—no fluff, pure instruction) and Progressive Disclosure Architecture (28/30—excellent file layering). The weakness is Spec Compliance (11/15)—mostly because your frontmatter description lacks trigger phrases, which actually costs you discoverability.

What's Working Well

• Copy-paste checklists are gold. The validation checklist at lines 251–263 gives users actual gates to pass before delegating. That's the kind of thing people bookmark.
• Three specialist templates (Architecture, Mentor, ML-Architect) show real flexibility without bloat. Each one is complete but brief enough to customize.
• The CONTEXT/TASK/REQUIREMENTS structure is clean. You're teaching a mental model, not just listing steps. That's why the Writing Style score is maxed out.
• File organization hits the mark—270 lines in the main SKILL.md, templates tucked into /references/. Token-efficient without hiding anything.

The Big One: Frontmatter Description Needs Trigger Phrases

Your description currently reads: "Transform clarified user requests into structured delegation prompts optimized for specialist agents..."

This doesn't have trigger phrases. According to the spec, it should look like:

description: Performs delegation prompt crafter operations. Use when asked to "delegation prompt crafter", "run delegation prompt crafter", "create delegation prompt", or "delegation prompt crafter help".

Why this matters: Without triggers, your skill becomes invisible to agent routing systems. Users won't know how to ask for it. You lose 2+ points on spec compliance, but more importantly, you lose adoption. This is a 5-minute fix that pushes you from 99 to 100+.

Other Things Worth Fixing

1. Add integration flow example – Lines 12–16 mention routing to mentor for validation, but don't show the full chain: clarification → architect → mentor → implementation. A simple flowchart or numbered sequence would close that gap. (+2 points potential)

2. Validation checklist needs measurability – "Task is specific with clear deliverable" is binary but subjective. Add guardrails: (one sentence, specific to use case, not reusable template). Helps users know when they've actually passed the gate. (+2 points)

3. Add "When NOT to Use" section – You cover when to use this, but not when it's overkill (simple Q&A, internal clarifications, quick routing). This prevents misuse and shows thoughtful design. (+1 point)

Quick Wins

• Update frontmatter description with trigger phrases (5 min, +2 points)
• Add workflow diagram showing full delegation chain (10 min, +2 points)
• Make validation checklist measurable with concrete criteria (5 min, +2 points)
• Add "When NOT to Use" section (3 min, +1 point)

These four fixes get you to 104+, but more importantly, they make this skill actually discoverable and bulletproof for edge cases.

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.