Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Use when you need to generate Architecture Decision Records (ADRs) for a Java project through an interactive, conversational process that systematically gathers context, stakeholders, options, and outcomes to produce well-structured ADR documents. This should trigger for requests such as Generate ADR; Create Architecture Decision Record; Document architecture decision; Architecture Decision Record for Java. Part of Plinth Toolkit
.claude/skills/jabrena-030-architecture-adr-general/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -34% | 0% |
| case-17 | ✗→✓ | ▲ Improved | -31% | 0% |
| case-11 | ✓→✗ | ▼ Worse | -32% | 0% |
| case-09 | ✓→✓ | ＝ Same ✓ | 94% | 0% |
| case-04 | ✓→✓ | ＝ Same ✓ | 35% | 0% |
Generate Architecture Decision Records (ADRs) for Java projects through an interactive, conversational process that elicits the decision inputs needed for a well-structured ADR. This is an interactive SKILL. Use only the current conversation and repository files explicitly available in the current session; do not expose, persist, or reuse unrelated user/project context.
What is covered in this Skill?
Handle ambiguity and blockers explicitly to avoid implicit assumptions.
Load references/030-architecture-adr-general.md from this skill, then elicit stakeholders, decision drivers, options, and trade-offs through the current conversation only.
Summarize recommended option, rationale, and consequences, and confirm alignment with the user before creating the ADR artifact.
Create a MADR-style ADR document with the final decision, alternatives, consequences, and follow-up actions.
For detailed guidance, examples, and constraints, see references/030-architecture-adr-general.md.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-09 | pass→pass | 3,938 | 4,020 | +2% | 1 | 1 | 0% | 565 | 1,098 | +94% | 0 | 0 | — |
case-01 | fail→pass | 12,353 | 5,533 | -55% | 1 | 1 | 0% | 2,020 | 1,343 | -34% | 0 | 0 | — |
case-02 | fail→fail | 16,425 | 19,885 | +21% | 1 | 1 | 0% | 3,000 | 3,714 | +24% | 0 | 0 | — |
case-03 | fail→fail | 18,222 | 17,710 | -3% | 1 | 1 | 0% | 2,827 | 2,039 | -28% | 0 | 0 | — |
case-04 | pass→pass | 5,291 | 4,413 | -17% | 1 | 1 | 0% | 899 | 1,210 | +35% | 0 | 0 | — |
case-05 | pass→pass | 8,295 | 7,086 | -15% | 1 | 1 | 0% | 1,379 | 1,596 | +16% | 0 | 0 | — |
case-06 | fail→fail | 14,310 | 13,949 | -3% | 1 | 1 | 0% | 2,247 | 2,609 | +16% | 0 | 0 | — |
case-07 | pass→pass | 8,390 | 4,412 | -47% | 1 | 1 | 0% | 1,199 | 1,178 | -2% | 0 | 0 | — |
case-08 | pass→pass | 5,843 | 4,001 | -32% | 1 | 1 | 0% | 953 | 1,095 | +15% | 0 | 0 | — |
case-10 | fail→fail | 11,269 | 13,098 | +16% | 1 | 1 | 0% | 2,019 | 2,523 | +25% | 0 | 0 | — |
case-11 | pass→fail | 14,569 | 7,457 | -49% | 1 | 1 | 0% | 2,328 | 1,580 | -32% | 0 | 0 | — |
case-12 | pass→pass | 4,080 | 2,802 | -31% | 1 | 1 | 0% | 512 | 918 | +79% | 0 | 0 | — |
case-13 | pass→pass | 15,263 | 7,965 | -48% | 1 | 1 | 0% | 2,452 | 1,908 | -22% | 0 | 0 | — |
case-14 | pass→pass | 12,827 | 5,668 | -56% | 1 | 1 | 0% | 1,979 | 1,324 | -33% | 0 | 0 | — |
case-15 | pass→pass | 10,314 | 7,027 | -32% | 1 | 1 | 0% | 1,521 | 1,691 | +11% | 0 | 0 | — |
case-16 | fail→fail | 13,213 | 10,434 | -21% | 1 | 1 | 0% | 2,154 | 2,194 | +2% | 0 | 0 | — |
case-17 | fail→pass | 15,969 | 12,425 | -22% | 1 | 1 | 0% | 2,150 | 1,484 | -31% | 0 | 0 | — |
case-18 | pass→pass | 7,782 | 5,041 | -35% | 1 | 1 | 0% | 1,317 | 1,200 | -9% | 0 | 0 | — |
case-19 | pass→pass | 8,053 | 6,402 | -21% | 1 | 1 | 0% | 1,396 | 1,437 | +3% | 0 | 0 | — |
case-20 | pass→pass | 16,104 | 9,415 | -42% | 1 | 1 | 0% | 3,204 | 2,320 | -28% | 0 | 0 | — |
case-21 | pass→pass | 8,824 | 14,639 | +66% | 1 | 1 | 0% | 1,480 | 1,608 | +9% | 0 | 0 | — |
case-22 | pass→pass | 5,855 | 4,329 | -26% | 1 | 1 | 0% | 941 | 1,209 | +28% | 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. The headline lift of +5 percentage points is the difference between those two pass rates over the 22 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.