Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Use this skill when the user asks to update CLAUDE.md, save a lesson, or persist something from the current session: phrases like "update claude.md", "what should we remember", "save this lesson", or "add to context". Scans the conversation for hard-won lessons, new file paths, infrastructure changes, and new workflows, then proposes scoped edits to the project's CLAUDE.md for approval before writing.
.claude/skills/jamditis-claude-md-updater/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 38% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 235% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 74% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 52% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 49% | 0% |
Analyze the current conversation to identify information worth persisting in CLAUDE.md for future sessions, then propose the edits for approval before writing anything.
Show the exact patch first, then a short summary grouped by category:
`## Proposed CLAUDE.md updates
--- a/CLAUDE.md +++ b/CLAUDE.md @@ section being changed @@ existing context line +new durable lesson, infra fact, or workflow
Summary of the diff above:
- Hard-won lessons: [one line, if any]
- Infrastructure updates: [one line, if any]
- New workflows: [one line, if any]
Apply this diff? Nothing is written until you approve.CLAUDE.md is team-shared (checked into git) and advisory: it loads into every session, so it is for durable, infrequently-changing facts, not a running log. Per the project-memory skill, keep these out of CLAUDE.md:
Route those to auto memory (~/.claude/projects/<project>/memory/, which Claude maintains on its own) or to a gitignored CLAUDE.local.md for personal notes. Putting them in CLAUDE.md bloats shared context and loads stale per-session facts into future sessions.
CLAUDE.local.md, never the committed CLAUDE.md.| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 8,627 | 6,463 | -25% | 1 | 1 | 0% | 1,509 | 2,077 | +38% | 0 | 0 | — |
case-02 | fail→fail | 10,339 | 8,354 | -19% | 1 | 1 | 0% | 1,698 | 2,355 | +39% | 0 | 0 | — |
case-03 | fail→fail | 8,923 | 32,688 | +266% | 1 | 1 | 0% | 1,571 | 3,723 | +137% | 0 | 0 | — |
case-04 | fail→fail | 5,761 | 4,795 | -17% | 1 | 1 | 0% | 1,016 | 1,835 | +81% | 0 | 0 | — |
case-05 | pass→pass | 9,016 | 7,884 | -13% | 1 | 1 | 0% | 1,430 | 2,288 | +60% | 0 | 0 | — |
case-06 | pass→pass | 13,828 | 8,388 | -39% | 1 | 1 | 0% | 2,002 | 2,257 | +13% | 0 | 0 | — |
case-07 | fail→pass | 2,660 | 3,488 | +31% | 1 | 1 | 0% | 465 | 1,556 | +235% | 0 | 0 | — |
case-08 | fail→pass | 6,952 | 4,971 | -28% | 1 | 1 | 0% | 1,034 | 1,800 | +74% | 0 | 0 | — |
case-09 | fail→pass | 5,309 | 3,711 | -30% | 1 | 1 | 0% | 1,011 | 1,541 | +52% | 0 | 0 | — |
case-10 | fail→pass | 5,764 | 3,988 | -31% | 1 | 1 | 0% | 1,053 | 1,570 | +49% | 0 | 0 | — |
case-11 | fail→pass | 8,093 | 3,361 | -58% | 1 | 1 | 0% | 1,532 | 1,412 | -8% | 0 | 0 | — |
case-12 | fail→pass | 9,451 | 6,518 | -31% | 1 | 1 | 0% | 1,555 | 2,089 | +34% | 0 | 0 | — |
case-13 | pass→pass | 4,776 | 3,718 | -22% | 1 | 1 | 0% | 973 | 1,554 | +60% | 0 | 0 | — |
case-14 | fail→pass | 4,521 | 8,470 | +87% | 1 | 1 | 0% | 870 | 2,536 | +191% | 0 | 0 | — |
case-15 | fail→pass | 5,015 | 4,181 | -17% | 1 | 1 | 0% | 846 | 1,538 | +82% | 0 | 0 | — |
case-16 | pass→pass | 10,539 | 2,596 | -75% | 1 | 1 | 0% | 1,709 | 1,305 | -24% | 0 | 0 | — |
case-17 | pass→pass | 10,170 | 3,708 | -64% | 1 | 1 | 0% | 1,810 | 1,643 | -9% | 0 | 0 | — |
case-18 | pass→pass | 4,076 | 4,408 | +8% | 1 | 1 | 0% | 729 | 1,659 | +128% | 0 | 0 | — |
case-19 | fail→pass | 5,879 | 4,461 | -24% | 1 | 1 | 0% | 1,131 | 1,698 | +50% | 0 | 0 | — |
case-20 | fail→pass | 8,773 | 5,986 | -32% | 1 | 1 | 0% | 1,542 | 1,949 | +26% | 0 | 0 | — |
case-21 | fail→pass | 6,381 | 4,163 | -35% | 1 | 1 | 0% | 1,183 | 1,597 | +35% | 0 | 0 | — |
case-22 | fail→pass | 8,564 | 6,668 | -22% | 1 | 1 | 0% | 1,498 | 2,058 | +37% | 0 | 0 | — |
case-23 | pass→pass | 4,570 | 2,962 | -35% | 1 | 1 | 0% | 872 | 1,354 | +55% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 23 cases were attempted. The headline lift of +57 percentage points is the difference between those two pass rates over the 23 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.