Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Verify every @path chain import and every markdown link inside every CLAUDE.md in this project resolves to an existing file. Read-only — returns broken links with file:line refs, never edits.
.claude/skills/alirezarezvani-claude-md-link-check/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-05 | ✗→✓ | ▲ Improved | 53% | 0% |
| case-14 | ✓→✗ | ▼ Worse | -35% | 0% |
| case-04 | ✓→✓ | ＝ Same ✓ | 82% | 0% |
| case-15 | ✓→✓ | ＝ Same ✓ | 33% | 0% |
| case-08 | ✓→✓ | ＝ Same ✓ | -30% | 0% |
Optional path-glob: $ARGUMENTS (default . — entire tree).
Run these steps in order. Do not modify any file.
find <root> -name "CLAUDE.md" -type f -not -path "*/.git/*" -not -path "*/node_modules/*". Also include .claude/rules/*.md. Record paths.^@\S+. The literal after @ is a relative path.[text](target) where target is not an HTTP(S) URL, not mailto:, not a bare anchor #section.Read on the parent file to confirm position, then test -e <resolved-path> or Glob).@../CLAUDE.md inside skill/CLAUDE.md, resolved path is CLAUDE.md.[Backend](backend/CLAUDE.md) inside the root, resolved path is backend/CLAUDE.md.## Link Check
Files inspected: <count>
References checked: <chain_imports> @-imports, <md_links> markdown links
### Broken
- <file>:<line> — `<original-target>` → does not resolve (expected `<absolute-path>`)
- ... (omit section if empty)
### Clean
<count> references resolved.## Link Check\n\nAll <N> references resolved across <M> files.. Do not pad.Hard rule: never invent a fix. Report the broken target verbatim. Repair is the user's call (or /sync-claude-md's).
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 3,946 | 7,334 | +86% | 1 | 1 | 0% | 458 | 778 | +70% | 0 | 0 | — |
case-02 | fail→fail | 5,265 | 7,139 | +36% | 1 | 1 | 0% | 300 | 887 | +196% | 0 | 0 | — |
case-03 | fail→fail | 6,146 | 6,859 | +12% | 1 | 1 | 0% | 232 | 731 | +215% | 0 | 0 | — |
case-04 | pass→pass | 9,612 | 18,484 | +92% | 1 | 1 | 0% | 1,519 | 2,762 | +82% | 0 | 0 | — |
case-05 | fail→pass | 11,548 | 22,128 | +92% | 1 | 1 | 0% | 2,106 | 3,231 | +53% | 0 | 0 | — |
case-06 | fail→fail | 11,327 | 36,511 | +222% | 1 | 1 | 0% | 1,901 | 1,037 | -45% | 0 | 0 | — |
case-15 | pass→pass | 5,238 | 3,737 | -29% | 1 | 1 | 0% | 834 | 1,108 | +33% | 0 | 0 | — |
case-07 | fail→fail | 2,913 | 5,084 | +75% | 1 | 1 | 0% | 388 | 719 | +85% | 0 | 0 | — |
case-08 | pass→pass | 8,461 | 3,122 | -63% | 1 | 1 | 0% | 1,372 | 966 | -30% | 0 | 0 | — |
case-09 | pass→pass | 6,052 | 3,681 | -39% | 1 | 1 | 0% | 979 | 1,155 | +18% | 0 | 0 | — |
case-10 | pass→pass | 8,197 | 2,887 | -65% | 1 | 1 | 0% | 1,194 | 950 | -20% | 0 | 0 | — |
case-11 | pass→pass | 9,908 | 2,237 | -77% | 1 | 1 | 0% | 1,494 | 777 | -48% | 0 | 0 | — |
case-12 | pass→pass | 6,677 | 2,509 | -62% | 1 | 1 | 0% | 998 | 795 | -20% | 0 | 0 | — |
case-13 | pass→pass | 5,785 | 3,095 | -46% | 1 | 1 | 0% | 908 | 930 | +2% | 0 | 0 | — |
case-14 | pass→fail | 9,583 | 3,448 | -64% | 1 | 1 | 0% | 1,514 | 979 | -35% | 0 | 0 | — |
case-16 | pass→pass | 4,795 | 2,283 | -52% | 1 | 1 | 0% | 749 | 807 | +8% | 0 | 0 | — |
case-17 | pass→pass | 7,307 | 5,418 | -26% | 1 | 1 | 0% | 1,064 | 1,018 | -4% | 0 | 0 | — |
case-18 | pass→pass | 5,962 | 2,321 | -61% | 1 | 1 | 0% | 809 | 842 | +4% | 0 | 0 | — |
case-19 | fail→fail | 3,236 | 1,610 | -50% | 1 | 1 | 0% | 435 | 665 | +53% | 0 | 0 | — |
case-20 | pass→pass | 10,144 | 2,852 | -72% | 1 | 1 | 0% | 1,458 | 883 | -39% | 0 | 0 | — |
case-21 | pass→pass | 12,848 | 2,782 | -78% | 1 | 1 | 0% | 2,024 | 881 | -56% | 0 | 0 | — |
case-22 | pass→pass | 9,708 | 2,050 | -79% | 1 | 1 | 0% | 1,474 | 746 | -49% | 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. 22 cases were attempted, and 17 counted toward the lift figure. The other 5 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of 0 percentage points is the difference between those two pass rates over the 17 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.
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.