social-media-analyzer

Found your social-media-analyzer skill while reviewing some recent submissions, and I'm curious about the design choice to separate parsing from analysis—that's usually where folks either go too minimal or over-engineer. The 62/100 puts it in that interesting territory where the core idea is solid but there's a clear gap between concept and execution.

Links:

• GitHub
• Your agentic skill in the SkillzWave marketplace

TL;DR

You're at 62/100, solidly in D territory. Based on Anthropic's best practices for skill architecture. Strongest area is Spec Compliance (11/15)—your frontmatter is clean and properly structured. Weakest is Writing Style (5/10) and Progressive Disclosure Architecture (18/30)—the skill reads more like marketing copy than actionable instructions, and there's a lot of token-inefficient fluff that should be trimmed or moved to supporting files.

What's Working Well

• Solid frontmatter structure - Your YAML is valid and includes all required fields, which puts you ahead on the spec compliance front
• Real supporting files - Having calculate_metrics.py, analyze_performance.py, and reference files shows you thought about modularity
• Specific problem domain - Social media analytics is concrete enough that the skill has real utility potential

The Big One: Missing Trigger Phrases

Here's what's costing you the most points: your description doesn't include trigger phrases. Claude won't know when to use this skill because you haven't told it.

Current description:

Analyzes social media campaign performance across platforms with engagement metrics, ROI calculations...

It needs to be:

Performs social media analysis and campaign performance tracking. Use when asked to 'analyze social media', 'calculate engagement rate', 'measure ROI', 'compare platform performance', 'track campaign metrics', or 'benchmark social media'.

This one change alone should net you +2 points and make Claude actually recognize when to invoke your skill. Right now it's invisible.

Other Things Worth Fixing

1. Trim the token bloat - Your Capabilities, Best Practices, and Limitations sections are verbose and belong in HOW_TO_USE.md, not SKILL.md. SKILL.md should be a execution guide, not a feature brochure. Moving this stuff could save ~40 tokens and bump PDA by +4 points.

2. Add numbered workflow steps - Instead of three example queries, give Claude a clear process: (1) Validate input completeness, (2) Run metric calculations, (3) Run performance analysis, (4) Generate insights report, (5) Verify against benchmarks. This gives +3 points on Ease of Use.

3. Add validation checks - Your skill is missing data quality safeguards. Add a section like: "Before analysis: check that reach > 0 (avoid division errors), date ranges are valid, all required metrics present. After analysis: verify engagement < 100%, ROI matches ad spend." This adds real utility and +2 points.

4. Reference your example files - You have sample_input.json and expected_output.json sitting unused. Add an Examples section that actually points to them. Makes it discoverable and adds +2 points.

Quick Wins

• Add trigger phrases to description (+2 points)
• Move verbose sections to supporting files (+4 points)
• Add numbered workflow steps (+3 points)
• Include validation and verification steps (+2 points)
• Reference example files with actual output (+2 points)

That's +13 points with straightforward structural fixes—gets you from 62 to 75 without changing the core idea.

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.