|
| 1 | +name: ADR (Architecture Decision Record) |
| 2 | +description: Propose a change to how sql-sop works (parser backend, distribution format, new rule pack, etc.) |
| 3 | +title: "[ADR] <one-line summary>" |
| 4 | +labels: ["ADR", "discussion"] |
| 5 | +body: |
| 6 | + - type: markdown |
| 7 | + attributes: |
| 8 | + value: | |
| 9 | + Use this template for proposals that change *how* sql-sop works, |
| 10 | + not just *what* it checks. Examples: new parser backend, |
| 11 | + new distribution format, new rule namespace (DBT*, contracts |
| 12 | + pack), new CLI flag. Simple rule additions belong in the |
| 13 | + normal "Rule request" template instead. |
| 14 | +
|
| 15 | + ADRs follow the pattern in `GOVERNANCE.md` § Architecture |
| 16 | + Decision Records. After discussion lands on a decision, the |
| 17 | + ADR issue stays open in an archived state as the record of |
| 18 | + "why we did it this way." A follow-up issue with the task |
| 19 | + breakdown links back to it. |
| 20 | +
|
| 21 | + - type: textarea |
| 22 | + id: what |
| 23 | + attributes: |
| 24 | + label: What |
| 25 | + description: One paragraph. What the proposal adds or changes. |
| 26 | + validations: |
| 27 | + required: true |
| 28 | + |
| 29 | + - type: textarea |
| 30 | + id: why |
| 31 | + attributes: |
| 32 | + label: Why |
| 33 | + description: | |
| 34 | + What gap this fills. What problem users hit today. Cite real |
| 35 | + examples if you have them. Skip the marketing. |
| 36 | + validations: |
| 37 | + required: true |
| 38 | + |
| 39 | + - type: textarea |
| 40 | + id: scope-in |
| 41 | + attributes: |
| 42 | + label: In scope |
| 43 | + description: What this PR / direction ships. Bulleted list. |
| 44 | + validations: |
| 45 | + required: true |
| 46 | + |
| 47 | + - type: textarea |
| 48 | + id: scope-out |
| 49 | + attributes: |
| 50 | + label: Out of scope (or deferred) |
| 51 | + description: | |
| 52 | + What this proposal explicitly does not cover. Per |
| 53 | + `GOVERNANCE.md` hard lines, "rule-based not AI" and "stable |
| 54 | + rule IDs" are always out unless this proposal is specifically |
| 55 | + about changing them. |
| 56 | + validations: |
| 57 | + required: true |
| 58 | + |
| 59 | + - type: textarea |
| 60 | + id: compatibility |
| 61 | + attributes: |
| 62 | + label: Compatibility |
| 63 | + description: | |
| 64 | + What changes for existing users. Default behaviour, rule-ID |
| 65 | + impact, new dependencies. Anything that breaks downstream |
| 66 | + configs needs a deprecation plan -- see `GOVERNANCE.md`. |
| 67 | + validations: |
| 68 | + required: true |
| 69 | + |
| 70 | + - type: textarea |
| 71 | + id: plan |
| 72 | + attributes: |
| 73 | + label: Implementation plan |
| 74 | + description: | |
| 75 | + Numbered list of the actual work. Files touched, new tests, |
| 76 | + README/CHANGELOG updates. Use this to estimate effort and |
| 77 | + catch scope creep early. |
| 78 | + validations: |
| 79 | + required: false |
| 80 | + |
| 81 | + - type: textarea |
| 82 | + id: open-questions |
| 83 | + attributes: |
| 84 | + label: Open questions |
| 85 | + description: | |
| 86 | + Decisions not yet made. State your lean for each so the |
| 87 | + discussion has a starting point. Format: "**Question.** Lean: x." |
| 88 | + validations: |
| 89 | + required: false |
| 90 | + |
| 91 | + - type: textarea |
| 92 | + id: rejected |
| 93 | + attributes: |
| 94 | + label: What I rejected |
| 95 | + description: | |
| 96 | + Alternatives considered and dismissed, with one-sentence |
| 97 | + reasoning each. Documents the why so reviewers don't have to |
| 98 | + re-litigate the same ground. |
| 99 | + validations: |
| 100 | + required: false |
| 101 | + |
| 102 | + - type: checkboxes |
| 103 | + id: hard-lines |
| 104 | + attributes: |
| 105 | + label: Hard-lines check |
| 106 | + description: Tick what applies; helps reviewers triage faster. |
| 107 | + options: |
| 108 | + - label: This proposal stays inside the "rule-based, not AI" hard line |
| 109 | + - label: This proposal does not rename or remove an existing rule ID |
| 110 | + - label: This proposal preserves the low-false-positive-rate goal |
| 111 | + - label: I'm willing to implement this myself if it lands |
0 commit comments