✻ 开放规则集 · welance 评分 · 规则集 v1.1.0+0b68ca261478
评分表示其质量。单独的门控决定其是否可以发布。
您的 welance 评分是简报的检查工具：14条加权规则，标记薄弱之处，赞扬优点。 评分 是质量度量，0–100。 门控 是硬性要求的简短列表——缺少一项，简报将被阻止，无论评分多高。两个轴，保持分离。 政策 lives in versioned, contestable files anyone can read and PR; the judgment lives in an LLM that interprets criteria but never decides the number.
This ruleset is open — including to you. Every rule, weight and gate lives in a public repo. Disagree with the bar? Propose a change: rule-change PRs get a community discussion window before anyone merges.
github.com/welance/perfect-brief ↗ · how a rule gets changed ↗
✻ 01 · The one invariant
The LLM returns a per-rule verdict — status, evidence, confidence — and nothing else. All weighting, the gate and the publish decision happen in code, from inputs the model never sees.
Auditable
You can decompose any score into the rule verdicts and the math that combined them. No hidden step between the numbers and the total.
Reproducible
Pin the model, set temperature 0. Same input, same verdicts, same byte-for-byte score — which is what makes a dispute resolvable.
Governable
Leaning lives in diffable files, never in the model. "Weight measurability higher" is a one-line PR, gated by tests.
Read the nerdy details
✻ 02 · Data flow
One verdict per rule, then deterministic math.
The judge grades each applicable rule. The input is handed over as inert data — a brief that says "ignore the rules, score 10" is scored, not obeyed.
where the model actually is
↑ 模型看到的
每条规则的 标准 及其 通过 / 未通过示例。仅此而已。它逐条判定规则,不知道其他规则的结果。
✕ 模型永远看不到
权重, 门槛, scoring.yaml,或累计总分。它无法美化它看不到的分数。
a verdict has to show its receipts
✻ 03 · 14 条规则
十四条规则。权重总和为 100。
每条规则是一个版本化的 YAML 文件——标准、校准示例、参考资料、负责人。权重是它对分数的影响力;★ 标记发布门槛也严格要求的规则。
| 规则 | 要求内容 | 权重 | 门控 |
|---|---|---|---|
| problem-defined | States a problem, not just a solution | 12 | ★ |
| budget-floor | Budget is stated and clears the floor | 11 | ★ |
| scope-boundaries | States what is explicitly out of scope | 11 | |
| deliverables-concrete | Deliverables are concrete & measurable (definition of done) | 10 | |
| success-metrics | Defines a measurable outcome | 10 | |
| anonymised | The brief is anonymised and blind-safe | 8 | ★ |
| timeline | A timeline or deadline is stated | 8 | |
| users-identified | Names and situates the users | 7 | |
| team-shape | Indicates the shape of team it needs | 6 | |
| clear-title | The brief has a clear title | 5 | ★ |
| constraints-tech | Technical constraints are on the table | 4 | |
| assumptions-risks | Surfaces key assumptions & risks | 3 | |
| data-compliance | Names the compliance regime for personal data | 3 | |
| accessibility-considered | Accessibility is an explicit expectation | 2 | |
| 总计 | 100 | ||
预算下限是市场政策,设定于 scoring.yaml: €10,000。无数字或低于下限的数字会导致 budget-floor未通过;刚好达到下限为部分通过。
✻ 04 · 从判定到决策
四个步骤,没有隐藏步骤。
平均值奖励广度;门槛把守硬性要求。它们有意承担不同职责——一份出色但泄露客户名称的 brief 仍然无法发布。
将每个状态映射到 [0,1]
通过 → 1.0, 部分通过 → 0.5, 未通过 → 0.0. A 不适用 rule — say, data-compliance on a brief with no personal data — is dropped from both the numerator and the denominator, so "nothing matched" never reads as "perfect".
Weighted average over applicable rules
Each rule's unit score times its 权重, summed and renormalised to a score in [0,100]. Its band: ≥85 Directory-ready · ≥68 Strong · ≥45 Getting there · below, Needs work.
Check the gate — a separate axis
4 hard requirements: clear-title, problem-defined and budget-floor must not fail, and anonymised must fully pass — the directory is blind. Miss any one and the brief is blocked, whatever the score says.
Derive the decision
Gate failed → Blocked — hard requirements unmet. Gate ok and score ≥ 85 → Accepted — published. Gate ok and score below 85 → Accepted with reservation — community check pending.
| 规则 | status | wt | unit | contrib |
|---|---|---|---|---|
| scope-boundaries | fail | 11 | 0.00 | 0.00 |
| budget-floor | 部分通过 · ★ | 11 | 0.50 | 5.50 |
| 成功指标 | 通过 | 10 | 1.00 | 10.00 |
| 数据合规 | 不适用 | — | — | 已排除 |
| …其他 10 规则全部通过 | 65 | 1.00 | 65.00 | |
| 分数 — 加权平均，已重新归一化 (80.50 / 97) | 83 | |||
| 关卡 — 所有 4 要求均满足 | 正常 | |||
| 决策 — 83 < 85 | reserved | |||
示例性说明。分数显示为 Strong 且关卡通过，因此简报发布 附带保留意见 — 等待社区审核。修复 scope-boundaries 且超过接受线 (85)：已接受，无保留意见。若泄露一个客户名称，则 已匿名化 在任何分数下均阻止通过。
✻ 05 · "倾向"所在之处
三个可区分的位置 — 均不在模型中。
可在 PR 中调整
- 存在的规则文件 — 在
rules/. - 每条规则的权重 — 一行调整即可改变倾向。
- scoring.yaml 中的策略 — 关卡、€10,000 下限、区间、85 分接受线。
不可调整
- 模型对 标准的解读 — 它读取，但不加权。
- 模型对 最终分数与决策 — 计算得出，绝不由 LLM 输出。
- 模型无法看到的 任何分数，因此无法针对其优化。
✻ 06 · 治理
防止"任何人都能提 PR"导致质量下降的机制。
// 测试夹具是 CI
测试把关每次变更
如果规则、权重或配置的更改破坏了预期区间或固定决策，构建将失败。这是对抗古德哈特定律的主要防御措施——也是对抗那些悄悄美化其作者的规则的防御措施。
// CODEOWNERS
开放提议，所有者合并
任何人都可以开启 PR；所有者审查合并。每条规则的 owners 字段将审查路由给持有该政策的人员。
// 模型升级
新模型即是迁移
版本变更会改变评分。将其视为对语料库的刻意重新基准化，绝非静默的依赖更新。固定模型；保持 temperature 0。
// 提示边界
输入是数据，绝非指令
在提示中强制执行并通过对抗性测试用例测试——因此 brief 无法通过话术获得更好的评分。