Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Turns a work item — feature, story, bug fix, change request — into working code, reviewed and verified. Use when the user hands over an outcome and leaves the edits to you; a bare story or issue link counts. Also use whenever the user asks BMAD by name — then any change qualifies, even a tiny fully-specified edit. Do not volunteer for interactive edits the user directs and reviews themselves, or for version-control operations that record existing work without changing it.
.claude/skills/bmad-code-org-bmad-build/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-04 | ✗→✓ | ▲ Improved | -57% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -14% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -76% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 18% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -47% | 0% |
Run the following command exactly once without changing the current working directory. Replace {project-root} with the absolute path to the project root and {skill-root} with the absolute path to this skill's directory:
bashuv run --no-cache "{project-root}/_bmad/scripts/render_skill.py" --project-root "{project-root}" --skill "{skill-root}"
oneshot or full), append --set workflow.route=<value> to the command.none, quick, or thorough; "skip review" or "no review" mean none), append --set workflow.review=<value> to the command.workflow.md instruction printed to stdout.bmad skill's setup, installing bmad first if you do not have it (npx skills add bmad-code-org/BMAD-METHOD --skill bmad), then run the command above once more.uv being unavailable), report the command output and HALT. Do not run any workflow source directly.| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 8,615 | 7,435 | -14% | 1 | 1 | 0% | 777 | 825 | +6% | 0 | 0 | — |
case-02 | fail→fail | 5,825 | 7,500 | +29% | 1 | 1 | 0% | 380 | 621 | +63% | 0 | 0 | — |
case-03 | fail→fail | 13,862 | 11,405 | -18% | 1 | 1 | 0% | 2,041 | 996 | -51% | 0 | 0 | — |
case-04 | fail→pass | 26,762 | 10,492 | -61% | 1 | 1 | 0% | 3,339 | 1,439 | -57% | 0 | 0 | — |
case-05 | fail→pass | 13,685 | 5,401 | -61% | 1 | 1 | 0% | 1,308 | 1,127 | -14% | 0 | 0 | — |
case-06 | fail→pass | 18,330 | 3,930 | -79% | 1 | 1 | 0% | 2,900 | 707 | -76% | 0 | 0 | — |
case-07 | fail→pass | 9,000 | 9,328 | +4% | 1 | 1 | 0% | 1,496 | 1,770 | +18% | 0 | 0 | — |
case-08 | fail→pass | 10,759 | 3,845 | -64% | 1 | 1 | 0% | 1,604 | 843 | -47% | 0 | 0 | — |
case-09 | fail→pass | 9,683 | 3,614 | -63% | 1 | 1 | 0% | 1,201 | 738 | -39% | 0 | 0 | — |
case-10 | pass→pass | 12,251 | 18,129 | +48% | 1 | 1 | 0% | 1,097 | 2,326 | +112% | 0 | 0 | — |
case-11 | pass→pass | 6,904 | 3,856 | -44% | 1 | 1 | 0% | 963 | 769 | -20% | 0 | 0 | — |
case-12 | fail→fail | 24,600 | 11,291 | -54% | 1 | 1 | 0% | 2,627 | 2,000 | -24% | 0 | 0 | — |
case-13 | fail→fail | 6,411 | 2,765 | -57% | 1 | 1 | 0% | 827 | 627 | -24% | 0 | 0 | — |
case-14 | fail→pass | 9,994 | 9,281 | -7% | 1 | 1 | 0% | 732 | 591 | -19% | 0 | 0 | — |
case-15 | fail→fail | 8,651 | 8,920 | +3% | 1 | 1 | 0% | 1,322 | 812 | -39% | 0 | 0 | — |
case-16 | fail→pass | 23,538 | 3,133 | -87% | 1 | 1 | 0% | 1,303 | 692 | -47% | 0 | 0 | — |
case-17 | fail→pass | 76,041 | 3,474 | -95% | 1 | 1 | 0% | 2,098 | 546 | -74% | 0 | 0 | — |
case-18 | pass→pass | 47,319 | 2,711 | -94% | 1 | 1 | 0% | 1,783 | 548 | -69% | 0 | 0 | — |
case-19 | fail→pass | 28,535 | 2,719 | -90% | 1 | 1 | 0% | 1,905 | 622 | -67% | 0 | 0 | — |
case-20 | pass→fail | 7,056 | 13,838 | +96% | 1 | 1 | 0% | 947 | 782 | -17% | 0 | 0 | — |
case-21 | pass→fail | 10,290 | 30,491 | +196% | 1 | 1 | 0% | 1,595 | 879 | -45% | 0 | 0 | — |
case-22 | pass→pass | 10,451 | 17,730 | +70% | 1 | 1 | 0% | 1,620 | 2,103 | +30% | 0 | 0 | — |
case-23 | pass→pass | 8,864 | 17,164 | +94% | 1 | 1 | 0% | 1,267 | 2,947 | +133% | 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, and 16 counted toward the lift figure. The other 7 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 +35 percentage points is the difference between those two pass rates over the 16 comparable cases. 2 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.