Skillsinsights-to-prompts
Turn a Claude Code /insights report into before/after rewrites of the prompts that caused friction — both the prompts you type in sessions and the prompt text inside your own skills.
5 files and 419 lines, free under MIT. Built from arm-pz/insights-to-prompts at 432793c.
Turn a Claude Code /insights report into before/after rewrites of the prompts that caused friction — both the prompts you type in sessions and the prompt text inside your own skills. Reads ~/.claude/usage-data (report.html, cached facets) plus session transcripts with a bundled script, ranks friction that better wording could have prevented, and rewrites the worst offenders.
Version, date and limits are read from the repo on every sync, and the check runs every night. Real-world runs are listed only with their evidence.
Measured from the packed archive at commit 432793c. A skill is plain text your agent reads, so this list is all of it.
| File | Lines | Bytes |
|---|---|---|
| SKILL.md | 52 | 4,349 B |
| references/digest-format.md | 13 | 1,279 B |
| references/rewrite-rules.md | 42 | 2,287 B |
| references/skill-prompts.md | 21 | 1,247 B |
| scripts/extract_insights.py | 291 | 12,068 B |
| 5 files | 419 | 8 KB |
Read it before you install it: the skill lives at . in arm-pz/insights-to-prompts.