Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Analyze Elixir/Phoenix technical debt — duplicates, refactoring opportunities, credo issues. Use when asked about code quality, cleanup, or what to improve.
.claude/skills/oliver-kriska-phx-techdebt/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-08 | ✗→✓ | ▲ Improved | -55% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 1% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -16% | 0% |
| case-11 | ✗→✓ | ▲ Improved | -13% | 0% |
| case-17 | ✗→✓ | ▲ Improved | -22% | 0% |
Find and eliminate duplicate code patterns, anti-patterns, and refactoring opportunities in Elixir/Phoenix projects.
Run mix credo --strict.
Focus on:
Use Grep to search for repeated Repo calls (Repo.get!, Repo.get, Repo.one) in lib/**/*.ex. Use Grep to find duplicate query patterns (from.*in.*where) in lib/**/*.ex.
Use Grep with output_mode: "count" to count def changeset occurrences in lib/**/*.ex. Use Grep to find repeated validations (validate_required, validate_format) in lib/**/*.ex.
Use Grep to find similar action patterns (def create, def update, def delete) in lib/*_web/**/*.ex.
| Pattern | Symptom | Solution | |---------|---------|----------| | Repeated queries | Same Repo.get in multiple contexts | Create shared query module | | Duplicate validations | Same validate_* calls | Extract to shared changeset | | Similar controllers | Copy-pasted CRUD actions | Use Phoenix generators consistently | | Repeated transforms | Same Enum.map patterns | Extract to domain module |
For each duplication found, report:
Run /skill:phx-techdebt to analyze the codebase and generate a prioritized report of technical debt with specific remediation steps.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 27,682 | 4,955 | -82% | 1 | 1 | 0% | 5,403 | 852 | -84% | 0 | 0 | — |
case-02 | fail→fail | 18,458 | 4,798 | -74% | 1 | 1 | 0% | 3,467 | 732 | -79% | 0 | 0 | — |
case-03 | pass→pass | 12,195 | 6,156 | -50% | 1 | 1 | 0% | 1,954 | 1,467 | -25% | 0 | 0 | — |
case-08 | fail→pass | 12,822 | 2,992 | -77% | 1 | 1 | 0% | 2,259 | 1,010 | -55% | 0 | 0 | — |
case-04 | pass→pass | 3,913 | 3,783 | -3% | 1 | 1 | 0% | 618 | 1,048 | +70% | 0 | 0 | — |
case-05 | pass→pass | 4,813 | 1,351 | -72% | 1 | 1 | 0% | 781 | 680 | -13% | 0 | 0 | — |
case-06 | fail→pass | 9,797 | 6,624 | -32% | 1 | 1 | 0% | 1,651 | 1,665 | +1% | 0 | 0 | — |
case-07 | pass→pass | 15,339 | 13,399 | -13% | 1 | 1 | 0% | 2,379 | 2,786 | +17% | 0 | 0 | — |
case-09 | fail→fail | 15,674 | 10,414 | -34% | 1 | 1 | 0% | 2,869 | 2,365 | -18% | 0 | 0 | — |
case-10 | fail→pass | 6,004 | 2,805 | -53% | 1 | 1 | 0% | 1,041 | 879 | -16% | 0 | 0 | — |
case-11 | fail→pass | 12,221 | 7,709 | -37% | 1 | 1 | 0% | 2,055 | 1,791 | -13% | 0 | 0 | — |
case-12 | pass→pass | 14,165 | 9,392 | -34% | 1 | 1 | 0% | 2,487 | 2,207 | -11% | 0 | 0 | — |
case-13 | fail→fail | 9,574 | 13,833 | +44% | 1 | 1 | 0% | 1,559 | 2,759 | +77% | 0 | 0 | — |
case-14 | pass→pass | 13,811 | 10,054 | -27% | 1 | 1 | 0% | 2,377 | 2,246 | -6% | 0 | 0 | — |
case-15 | pass→pass | 12,531 | 2,327 | -81% | 1 | 1 | 0% | 2,054 | 825 | -60% | 0 | 0 | — |
case-16 | pass→pass | 15,279 | 5,357 | -65% | 1 | 1 | 0% | 2,508 | 1,373 | -45% | 0 | 0 | — |
case-17 | fail→pass | 8,714 | 4,498 | -48% | 1 | 1 | 0% | 1,475 | 1,150 | -22% | 0 | 0 | — |
case-18 | pass→fail | 14,189 | 6,395 | -55% | 1 | 1 | 0% | 2,266 | 1,666 | -26% | 0 | 0 | — |
case-19 | pass→pass | 11,801 | 3,684 | -69% | 1 | 1 | 0% | 1,953 | 1,118 | -43% | 0 | 0 | — |
case-20 | pass→pass | 17,198 | 16,991 | -1% | 1 | 1 | 0% | 2,694 | 3,203 | +19% | 0 | 0 | — |
case-21 | pass→pass | 17,184 | 11,833 | -31% | 1 | 1 | 0% | 3,132 | 2,556 | -18% | 0 | 0 | — |
case-22 | pass→pass | 8,747 | 7,432 | -15% | 1 | 1 | 0% | 1,574 | 1,800 | +14% | 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 +18 percentage points is the difference between those two pass rates over the 21 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.