Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Verifies that a Home Assistant integration follows a specific quality scale rule, checking whether it implements the required patterns, configurations, or code structures defined by the quality scale system. Use when asked to check a rule (e.g. "check if the peblar integration follows the config-flow rule") or to verify an integration reaches a quality tier (Bronze, Silver, Gold, Platinum).
.claude/skills/home-assistant-ha-quality-scale-verify/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-07 | ✗→✓ | ▲ Improved | 0% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 51% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 14% | 0% |
| case-12 | ✗→✓ | ▲ Improved | -1% | 0% |
| case-16 | ✗→✓ | ▲ Improved | -24% | 0% |
You are verifying whether a Home Assistant integration follows a specific quality scale rule. Verify one rule at a time; to check a full tier, verify each of that tier's rules (run in parallel subagents when possible).
Retrieve the official rule documentation from: https://raw.githubusercontent.com/home-assistant/developers.home-assistant/refs/heads/master/docs/core/integration-quality-scale/rules/{rule_name}.md where {rule_name} is the rule identifier (e.g. config-flow, entity-unique-id, parallel-updates).
Parse the rule documentation to identify:
Examine the integration's codebase at homeassistant/components/<integration domain>, focusing on:
manifest.json for quality scale declaration and configurationquality_scale.yaml for rule status (done, todo, exempt)Additional sources:
https://raw.githubusercontent.com/home-assistant/home-assistant.io/refs/heads/current/source/_integrations/<integration domain>.markdownhttps://pypi.org/pypi/<package>/jsondone, todo, or exempt in quality_scale.yamlexempt, verify the exemption reason is validdone, verify the actual implementation matches the requirementsQuality scale rules are cumulative: Bronze rules apply to all integrations with a quality scale, Silver rules apply to Silver+ integrations, and so on. Always consider the integration's target quality level when determining which rules to enforce.
Report only the rules that have issues. Do not list rules that pass or that are validly exempt. For each rule with an issue, provide:
If no rules have issues, say so in a single line. Be thorough but focused: examine only the aspects relevant to the rules being verified. If you cannot access the rule documentation or find the integration code, clearly state what information is missing and what you would need to complete the verification.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 34,760 | 7,855 | -77% | 1 | 1 | 0% | 5,455 | 1,044 | -81% | 0 | 0 | — |
case-02 | fail→fail | 26,525 | 6,487 | -76% | 1 | 1 | 0% | 4,428 | 993 | -78% | 0 | 0 | — |
case-03 | fail→fail | 57,832 | 7,757 | -87% | 1 | 1 | 0% | 5,743 | 1,112 | -81% | 0 | 0 | — |
case-04 | pass→pass | 11,773 | 9,260 | -21% | 1 | 1 | 0% | 2,091 | 2,387 | +14% | 0 | 0 | — |
case-05 | pass→pass | 13,070 | 18,740 | +43% | 1 | 1 | 0% | 2,446 | 3,421 | +40% | 0 | 0 | — |
case-06 | pass→pass | 7,843 | 11,514 | +47% | 1 | 1 | 0% | 1,222 | 2,881 | +136% | 0 | 0 | — |
case-07 | fail→pass | 6,592 | 2,735 | -59% | 1 | 1 | 0% | 1,032 | 1,031 | -0% | 0 | 0 | — |
case-08 | fail→pass | 4,540 | 2,911 | -36% | 1 | 1 | 0% | 701 | 1,058 | +51% | 0 | 0 | — |
case-09 | pass→pass | 2,287 | 2,635 | +15% | 1 | 1 | 0% | 312 | 1,036 | +232% | 0 | 0 | — |
case-10 | fail→pass | 8,316 | 5,620 | -32% | 1 | 1 | 0% | 1,383 | 1,571 | +14% | 0 | 0 | — |
case-11 | pass→pass | 3,060 | 3,688 | +21% | 1 | 1 | 0% | 402 | 926 | +130% | 0 | 0 | — |
case-12 | fail→pass | 7,429 | 3,303 | -56% | 1 | 1 | 0% | 1,141 | 1,126 | -1% | 0 | 0 | — |
case-13 | pass→fail | 9,633 | 5,195 | -46% | 1 | 1 | 0% | 1,629 | 1,029 | -37% | 0 | 0 | — |
case-14 | pass→pass | 9,317 | 3,518 | -62% | 1 | 1 | 0% | 1,565 | 1,184 | -24% | 0 | 0 | — |
case-15 | pass→pass | 17,975 | 4,204 | -77% | 1 | 1 | 0% | 3,042 | 1,342 | -56% | 0 | 0 | — |
case-16 | fail→pass | 19,154 | 9,471 | -51% | 1 | 1 | 0% | 2,813 | 2,140 | -24% | 0 | 0 | — |
case-17 | pass→pass | 4,824 | 7,384 | +53% | 1 | 1 | 0% | 631 | 1,808 | +187% | 0 | 0 | — |
case-18 | pass→pass | 6,207 | 3,999 | -36% | 1 | 1 | 0% | 994 | 1,184 | +19% | 0 | 0 | — |
case-19 | pass→pass | 8,416 | 3,974 | -53% | 1 | 1 | 0% | 1,512 | 1,230 | -19% | 0 | 0 | — |
case-20 | fail→pass | 10,443 | 2,916 | -72% | 1 | 1 | 0% | 1,720 | 1,020 | -41% | 0 | 0 | — |
case-21 | fail→pass | 15,623 | 8,522 | -45% | 1 | 1 | 0% | 2,060 | 2,128 | +3% | 0 | 0 | — |
case-22 | pass→pass | 11,975 | 5,159 | -57% | 1 | 1 | 0% | 1,743 | 1,371 | -21% | 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 19 counted toward the lift figure. The other 3 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 +27 percentage points is the difference between those two pass rates over the 19 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.