feat(output-contract): add density gradient, symbol discipline, desensitization

- §3.1: output density gradient (minimal/standard/full by task count)
- §3 DO: multi-task grouping by deliverable, no decorative emoji
- §4.6: desensitization self-check (root_cause, scoring internals, gate fields)
- develop/SKILL.md: update resource navigation description (zh+en)

Context-Checkpoint: B

Author: Li-Sanze

skills/en/skills/sopify/develop/SKILL.md

@@ -25,7 +25,7 @@ description: Develop phase entry; routes task execution, state updates, KB sync,
 
 - Long rules: `references/develop-rules.md`
 - Shared writing standards: `../references/shared-writing-dna.md` (apply to all output)
-- Output contract: `../references/output-contract.md` (required sections, conditional enhancement, self-check)
+- Output contract: `../references/output-contract.md` (required sections, density gradient, symbol discipline, desensitization)
 - Output templates: `assets/*.md`
 - Task extraction script: `scripts/extract_pending_tasks.py`
 

skills/en/skills/sopify/references/output-contract.md

@@ -36,15 +36,31 @@ Do not force structure on every scenario. Decide whether and how to enhance base
 
 Constraint: use at most one primary structure per reply; avoid stacking table + tree + flow.
 
+### 3.1 Output Density Gradient
+
+Select output density based on scenario complexity. Avoid heavy output for simple tasks:
+
+| Density Level | Applicable Scenarios | Verification Summary |
+|--------------|---------------------|---------------------|
+| Minimal | quick-fix, single-file change, Q&A | Simplified table (≤3 columns), omit review conclusion |
+| Standard | 3-5 task regular develop | Table, omit zero-info columns |
+| Full | 6+ tasks or cross-module delivery | Full table + phased review conclusions |
+
+Density level is auto-inferred from task count and change scope; no user specification needed.
+
 **DO:**
 - Use a comparison table when multi-option tradeoffs are present — make differences visible at a glance
 - Simplify verification summary on success when all passed — omit columns with no information
 - Keep plain text when information shape does not match any structure
+- Trim output by density gradient: quick-fix should not produce a full verification table
+- Group by deliverable when completing 3+ independent tasks; merge same conclusions into one line
+- No decorative emoji in body text (verification summaries, review conclusions, Changes section)
 
 **DON'T:**
 - Multi-option tradeoffs without comparison structure = missed enhancement
 - Forcing tables on simple Q&A = over-enhancement
 - Mixing two or more primary structures in one reply = cognitive overload
+- Expanding 5 identical `passed` conclusions item-by-item = information dilution
 
 ## 4. Pre-output Self-check
 
@@ -54,3 +70,5 @@ Before outputting the final reply, verify:
 2. Status symbol correct: `✓` only when all verifications passed with no degradation or skips; otherwise `!`.
 3. Footer complete: `Changes:` + `Next:` must be present.
 4. Conditional enhancement applied: if the reply meets a §3 trigger condition, use the corresponding structured format.
+5. Density level matched: quick-fix should not produce a full verification table; complex delivery should not lack review conclusions.
+6. Desensitization: rewrite raw `root_cause` enums as human descriptions; do not expose internal scoring weights; do not display gate/handoff JSON fields (`resolution_id`, `run_stage`, etc.).

skills/zh/skills/sopify/develop/SKILL.md

@@ -25,7 +25,7 @@ description: 开发实施阶段入口；聚合任务执行、状态更新、知
 
 - 长规则：`references/develop-rules.md`
 - 共享写作规范：`../references/shared-writing-dna.md`（所有输出遵循）
-- 输出契约：`../references/output-contract.md`（必需 section、条件增强、自检）
+- 输出契约：`../references/output-contract.md`（必需 section、密度梯度、符号纪律、脱敏规则）
 - 输出模板：`assets/*.md`
 - 任务提取脚本：`scripts/extract_pending_tasks.py`
 

skills/zh/skills/sopify/references/output-contract.md

@@ -22,7 +22,7 @@
 
 表格列约束：只展示当前场景有信息量的列。列省略只影响最终展示，不影响内部验证判断。`reason_code` 是内部验证字段，不在用户面输出中展示。
 
-## 3. Conditional Enhancement & Format Selection
+## 3. 条件增强与格式选择
 
 不强制所有场景结构化。按信息形状和触发条件决定是否增强、用什么结构：
 
@@ -36,15 +36,31 @@
 
 约束：同一回复最多选一种主结构，避免表格 + 树 + 流程叠加。
 
+### 3.1 输出密度梯度
+
+按场景复杂度选择输出密度，避免简单任务输出过重：
+
+| 密度级别 | 适用场景 | 验证摘要形式 |
+|---------|---------|------------|
+| 精简 | quick-fix、单文件修改、咨询问答 | 简化表格（≤3 列），省略复审结论 |
+| 标准 | 3-5 任务的常规 develop | 表格，省略无信息量列 |
+| 完整 | 6+ 任务或跨模块交付 | 完整表格 + 分阶段复审结论 |
+
+密度级别由任务数和改动范围自动推断，不需要用户指定。
+
 **DO：**
 - 满足多方案取舍时用对比表，让差异一目了然
 - success 全部 passed 时简化验证摘要，省略无信息量列
 - 信息形状不匹配任何结构时保持纯文本
+- 按密度梯度裁剪输出：quick-fix 不输出完整验证表
+- 3+ 独立任务时按交付物分组输出，同类结论合并为一行
+- 正文（验证摘要、复审结论、Changes 区）不使用装饰性 emoji
 
 **DON'T：**
 - 满足多方案取舍却无对比结构 = 漏增强
 - 强制在简单问答上叠加表格 = 过度增强
 - 同一回复混用两种以上主结构 = 认知过载
+- 5 个任务结果逐项展开相同 `passed` 结论 = 信息稀释
 
 ## 4. 输出前自检
 
@@ -54,3 +70,5 @@
 2. 状态符是否正确：`✓` 仅当全部验证通过且无降级/跳过；否则必须 `!`。
 3. footer 是否完整：`Changes:` + `Next:` 必须存在。
 4. 条件增强是否触发：如当前回复满足 §3 增强触发条件，应使用对应结构化表达。
+5. 密度级别是否匹配：quick-fix 不应输出完整验证表；复杂交付不应缺少复审结论。
+6. 脱敏：`root_cause` 枚举值转人话描述；内部评分权重不外露；gate/handoff JSON 字段（`resolution_id`、`run_stage` 等）不展示。