LLM skill usage

Tracking my claude skill usage:

Top skills overall:

┌──────┬─────────────────────────────────────────────────────────┬────────┐ │ Rank │ Skill │ Count │ ├──────┼─────────────────────────────────────────────────────────┼────────┤ │ 1 │ grill-with-docs │ 44 │ ├──────┼─────────────────────────────────────────────────────────┼────────┤ │ 2 │ quarto-modify │ 23 │ ├──────┼─────────────────────────────────────────────────────────┼────────┤ │ 3 │ superpowers:writing-plans │ 20 │ ├──────┼─────────────────────────────────────────────────────────┼────────┤ │ 4 │ superpowers:subagent-driven-development │ 18 │ ├──────┼─────────────────────────────────────────────────────────┼────────┤ │ 5 │ superpowers:brainstorming │ 18 │ ├──────┼─────────────────────────────────────────────────────────┼────────┤ │ 6 │ grill-me │ 14 │ ├──────┼─────────────────────────────────────────────────────────┼────────┤ │ 7 │ quarto-render │ 10 │ ├──────┼─────────────────────────────────────────────────────────┼────────┤ │ 8 │ superpowers:finishing-a-development-branch │ 10 │ ├──────┼─────────────────────────────────────────────────────────┼────────┤ │ 9 │ write-a-skill │ 7 │ ├──────┼─────────────────────────────────────────────────────────┼────────┤ │ 10 │ verify / obsidian-markdown / firecrawl:firecrawl-search │ 7 each │ └──────┴─────────────────────────────────────────────────────────┴────────┘

By project:

Private repo names are redacted and replaced with a brief descriptive name

• PRIVATE-QUARTO-WEBSITE (129) — heaviest user; dominated by quarto-modify, grill-with-docs, superpowers:subagent-driven-development
• PRIVATE-CLINICAL-TRIALS-RESEARCH (43) — brainstorming, grill-with-docs, writing-plans
• PRIVATE-ETL (36) — grill-with-docs, writing-plans, grill-me
• obsidian-notes / vaults (22) — obsidian-markdown, write-a-skill, obsidian-bases
• Other projects: scattered, low usage

Observations:

• You lean heavily on the grilling skills (grill-with-docs + grill-me = 58) — design interrogation is a strong habit
• The superpowers planning/execution cluster accounts for ~70 invocations combined, suggesting consistent structured development workflows
• quarto-modify and quarto-render are project-specific to casebooks (all 23+10 from there)
• 22 skills only appear once, suggesting some were experimental or one-off