Analyze local agent logs to rank K-skill directories into remove, review, or keep groups.
npx skills add https://github.com/NomaDamas/k-skill --skill k-skill-cleaner
Please help me install this Agent Skill. Skill: k-skill-cleaner Source: https://github.com/NomaDamas/k-skill/tree/main/k-skill-cleaner Command: npx skills add https://github.com/NomaDamas/k-skill --skill k-skill-cleaner
K-skill bundle directories grow over time, and deleting the wrong one can break workflows; this Skill helps you identify never-used candidates using best-effort agent trigger evidence, while keeping deletions user-approved.
Ask the agent to run the k-skill-cleaner helper to scan your last 90 days of usage logs and output a ranked shortlist of K-skills to review or delete, without deleting anything automatically.
High-intent search queries and answers about installing and using this skill.
To find unused skills in a coding agent bundle, you can analyze local agent trigger logs and usage counts over a configurable time window to identify never-loaded candidates for safe removal.
Yes, you can automate cleanup planning for agent skills across platforms like Claude Code and Codex by scanning usage logs for trigger evidence and generating user-approved removal shortlists without automatic filesystem deletions.
The best way to analyze coding agent usage logs for dependency hygiene is to deterministically scan local JSONL files for skill mentions and SKILL.md load markers, then rank candidates into remove, review, and keep groups.
Yes, the skill removal process supports pre-aggregated usage data imports by allowing you to supply a pre-aggregated usage JSON file as an alternative to scanning local agent logs directly within the time window.
Limitations of automated skill bundle slimming include best-effort trigger counting that may miss usage evidence, a mandatory interview to protect kept skills, and a strict prohibition against performing filesystem deletions automatically.
seolcodingScan local coding agent logs to rank unused K-skills for cleanup.
lv7devRun verification commands and produce pass/fail status reports.
LaiTszKinExecute user-specified commands in a background terminal within a scheduled time window.
CNifeCount pi skill invocations from session logs and categorize usage tiers.
