Your skills folder is a graveyard. Measure it.
Find and cut the skills that never fire. Claude Code users with large skill/command collections.
what to steal
Track which skills actually fire, then delete the rest.
Our takeaway from the source, not a quote.
skillreaper-prune-unused-skills.toml
# Find and cut the skills that never fire # source: https://github.com/thousandflowers/skillreaper [rig] who = "Claude Code users with large skill/command collections" author = "thousandflowers (GitHub)" date = "2026-08-23" harness = ["Claude Code"] models = [] [keep] - skillreaper usage audit [cut] - Unused skills/commands
| thing | status | why (per source) |
|---|---|---|
| Unused skills/commands | flagged for removal | Author's own setup: "382 items loaded, 13 ever fired – 3% utilization", ~19,755 dead tokens per session (measured 2026-08-23) |
Source: https://github.com/thousandflowers/skillreaper · checked 2026-10-14
// similar rigs
~/rigs/paulund-herdr-cli.toml2026-10-05
Three desktop apps out, one terminal multiplexer in
steal → Use one session view for all CLIs instead of one app per harness.
CodexClaude CodeOpenCode Codex / Claude Code /
by Paulund
~/rigs/anthropic-cut-system-prompt.toml2026-07-26
Anthropic cut over 80% of a harness system prompt for newer models
steal → Re-read your system prompt after each model upgrade; newer models need less.
Claude Code Opus 5Fable 5 System prompt
by Anthropic (reported by onthewire.ai)
~/rigs/pycomet-dead-weight-audit.toml2026-07-22
103 capabilities installed, 90 never fired in 30 days
steal → Mine your session logs for what actually ran, then cut the rest.
Claude Code Haiku 90 of 103 installed ca
by pycomet