Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Analyzes changesets with risk scoring, categorization by type and impact, and release note preparation. Use when extracting insights from raw change data.
.claude/skills/athola-diff-analysis/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 94% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 184% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 117% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 120% | 0% |
| case-17 | ✗→✓ | ▲ Improved | 78% | 0% |
Structured method for analyzing changesets: categorize changes, assess risks, generate insights. Works for git diffs, configuration changes, API migrations, schema updates, or document revisions.
Trigger Keywords: diff, changes, release notes, changelog, migration, impact, risk assessment
Auto-Load When: Git diffs present, change analysis requested, impact assessment needed.
Load modules based on workflow stage:
modules/semantic-categorization.md for change categorization workflowmodules/risk-assessment-framework.md when risk assessment is neededmodules/git-diff-patterns.md when working with git repositoriessanctum:git-workspace-review for git data gatheringimbue:proof-of-work for capturing analysis evidenceimbue:structured-output for formatting final deliverablesdiff-analysis:baseline-establisheddiff-analysis:changes-categorizeddiff-analysis:risks-assesseddiff-analysis:summary-preparedMark each item complete as you finish the corresponding step.
diff-analysis:baseline-established)Define comparison scope: what states are being compared, boundary of analysis, and scale metrics.
For git contexts, load modules/git-diff-patterns.md. For other contexts, compare relevant artifacts.
diff-analysis:changes-categorized)Group changes by semantic type. Load modules/semantic-categorization.md for change categories, semantic categories, and prioritization.
diff-analysis:risks-assessed)Evaluate impact. Load modules/risk-assessment-framework.md for risk indicators, levels, and scoring methodology.
diff-analysis:summary-prepared)Synthesize findings: theme, scope with counts, risk level, review focus, dependencies. Format for downstream consumption (PR descriptions, release notes, reviews).
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 7,831 | 10,489 | +34% | 1 | 1 | 0% | 1,315 | 2,545 | +94% | 0 | 0 | — |
case-02 | fail→pass | 8,099 | 17,967 | +122% | 1 | 1 | 0% | 1,292 | 3,668 | +184% | 0 | 0 | — |
case-03 | fail→fail | 9,567 | 11,403 | +19% | 1 | 1 | 0% | 1,419 | 2,380 | +68% | 0 | 0 | — |
case-04 | fail→fail | 1,798 | 3,267 | +82% | 1 | 1 | 0% | 233 | 1,114 | +378% | 0 | 0 | — |
case-05 | pass→pass | 14,698 | 8,824 | -40% | 1 | 1 | 0% | 1,680 | 1,971 | +17% | 0 | 0 | — |
case-06 | fail→pass | 5,734 | 4,574 | -20% | 1 | 1 | 0% | 619 | 1,341 | +117% | 0 | 0 | — |
case-07 | pass→fail | 9,802 | 15,710 | +60% | 1 | 1 | 0% | 1,464 | 3,190 | +118% | 0 | 0 | — |
case-08 | pass→pass | 17,584 | 9,825 | -44% | 1 | 1 | 0% | 2,800 | 2,384 | -15% | 0 | 0 | — |
case-09 | pass→pass | 7,487 | 8,147 | +9% | 1 | 1 | 0% | 1,082 | 1,880 | +74% | 0 | 0 | — |
case-10 | fail→pass | 7,518 | 10,929 | +45% | 1 | 1 | 0% | 1,178 | 2,592 | +120% | 0 | 0 | — |
case-17 | fail→pass | 10,078 | 12,690 | +26% | 1 | 1 | 0% | 1,568 | 2,788 | +78% | 0 | 0 | — |
case-11 | fail→fail | 6,135 | 7,303 | +19% | 1 | 1 | 0% | 963 | 1,141 | +18% | 0 | 0 | — |
case-12 | fail→fail | 6,963 | 12,619 | +81% | 1 | 1 | 0% | 964 | 2,396 | +149% | 0 | 0 | — |
case-13 | fail→pass | 8,468 | 13,134 | +55% | 1 | 1 | 0% | 1,374 | 2,720 | +98% | 0 | 0 | — |
case-14 | fail→pass | 20,406 | 18,493 | -9% | 1 | 1 | 0% | 3,199 | 3,750 | +17% | 0 | 0 | — |
case-15 | pass→pass | 25,459 | 18,502 | -27% | 1 | 1 | 0% | 2,253 | 2,663 | +18% | 0 | 0 | — |
case-16 | pass→pass | 10,572 | 12,381 | +17% | 1 | 1 | 0% | 1,633 | 2,579 | +58% | 0 | 0 | — |
case-18 | pass→fail | 23,203 | 18,789 | -19% | 1 | 1 | 0% | 1,824 | 3,760 | +106% | 0 | 0 | — |
case-19 | pass→pass | 16,094 | 17,165 | +7% | 1 | 1 | 0% | 2,518 | 3,835 | +52% | 0 | 0 | — |
case-20 | pass→pass | 16,631 | 17,542 | +5% | 1 | 1 | 0% | 2,902 | 3,685 | +27% | 0 | 0 | — |
case-21 | pass→pass | 21,794 | 17,443 | -20% | 1 | 1 | 0% | 3,980 | 3,714 | -7% | 0 | 0 | — |
case-22 | pass→pass | 14,607 | 15,739 | +8% | 1 | 1 | 0% | 2,399 | 3,227 | +35% | 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 21 counted toward the lift figure. The other 1 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 +23 percentage points is the difference between those two pass rates over the 21 comparable cases. 3 cases got worse with the skill loaded, and they are 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.