speccore 5.43.0 → 5.50.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/.agents/skills/spec-analyze/SKILL.md +7 -0
- package/.agents/skills/spec-ask/SKILL.md +10 -0
- package/.agents/skills/spec-change/SKILL.md +7 -0
- package/.agents/skills/spec-dev/SKILL.md +7 -0
- package/.agents/skills/spec-doc2spec/SKILL.md +7 -0
- package/.agents/skills/spec-execute/SKILL.md +7 -0
- package/.agents/skills/spec-iteration-create/SKILL.md +7 -0
- package/.agents/skills/spec-plan/SKILL.md +7 -0
- package/.agents/skills/spec-spec2doc/SKILL.md +7 -0
- package/.agents/skills/spec-split/SKILL.md +7 -0
- package/.agents/skills/spec-task-create/SKILL.md +17 -8
- package/.agents/skills/speccore-router/SKILL.md +8 -0
- package/package.json +1 -1
|
@@ -1,3 +1,10 @@
|
|
|
1
|
+
---
|
|
2
|
+
name: spec-analyze
|
|
3
|
+
description: >
|
|
4
|
+
AI 需求分析引擎。按任务类型(feature/bugfix/review/test/docs/refactor/deploy/security/performance)
|
|
5
|
+
自动生成对应分析报告。Use when user says "分析" "审查" "安全审计" "性能分析" "测试用例".
|
|
6
|
+
allowed-tools: ["Bash", "Read", "Write", "Edit"]
|
|
7
|
+
---
|
|
1
8
|
# SpecCore Analyze — 需求分析器
|
|
2
9
|
|
|
3
10
|
> **你负责**: 读取需求文档 → 你自己分析 → CLI 写入 ANALYSIS.md。你自己就是分析引擎。
|
|
@@ -1,3 +1,13 @@
|
|
|
1
|
+
---
|
|
2
|
+
name: spec-ask
|
|
3
|
+
description: >
|
|
4
|
+
SpecCore 可执行编排引擎。处理意图识别 → CLI 执行 → AI 内容生成 → 写入。
|
|
5
|
+
Use when user says anything that doesn't match a specific skill route,
|
|
6
|
+
or complex multi-step requests involving "然后" "接着" "定时" "分批" "调度".
|
|
7
|
+
Always invoked via execute_command("speccore ask '...'").
|
|
8
|
+
allowed-tools: ["Bash", "Read", "Write", "Edit"]
|
|
9
|
+
disable-model-invocation: false
|
|
10
|
+
---
|
|
1
11
|
# SpecCore Ask — 可执行编排引擎 v4
|
|
2
12
|
|
|
3
13
|
> **你的身份**: 你就是宿主 AI。你的任务是:理解用户意图 → 调用 CLI --prompt → 自己生成内容 → 调 CLI --response 写入。
|
|
@@ -1,3 +1,10 @@
|
|
|
1
|
+
---
|
|
2
|
+
name: spec-doc2spec
|
|
3
|
+
description: >
|
|
4
|
+
文档导入。Word/Excel/CSV → Spec Markdown,支持图片提取和 Bug 列表解析。
|
|
5
|
+
Use when user says "导入文档" "Word" "Excel" "CSV" "Bug列表".
|
|
6
|
+
allowed-tools: ["Bash", "Read", "Write", "Edit"]
|
|
7
|
+
---
|
|
1
8
|
# SpecCore doc2spec — AI + Pandoc 双路验证
|
|
2
9
|
|
|
3
10
|
> **你负责**: Pandoc 机械转换 + 你自己交叉验证 = 零数据丢失。
|
|
@@ -1,3 +1,10 @@
|
|
|
1
|
+
---
|
|
2
|
+
name: spec-execute
|
|
3
|
+
description: >
|
|
4
|
+
任务执行引擎。读取 ANALYSIS.md → Prompt/Apply 模式生成代码。
|
|
5
|
+
Use when user says "开发" "执行" "写代码" "实现" "跑任务".
|
|
6
|
+
allowed-tools: ["Bash", "Read", "Write", "Edit"]
|
|
7
|
+
---
|
|
1
8
|
# SpecCore Execute — 代码生成执行器
|
|
2
9
|
|
|
3
10
|
> **你负责**: 读取 Task Spec → 构建代码生成上下文 → 你自己生成代码 → CLI 写入文件。
|
|
@@ -1,3 +1,10 @@
|
|
|
1
|
+
---
|
|
2
|
+
name: spec-iteration-create
|
|
3
|
+
description: >
|
|
4
|
+
迭代创建。AI 英文关键词命名 + 平台检查。
|
|
5
|
+
Use when user says "创建迭代" "新建迭代" "Sprint".
|
|
6
|
+
allowed-tools: ["Bash", "Read", "Write", "Edit"]
|
|
7
|
+
---
|
|
1
8
|
# SpecCore Iteration Create — 交互式迭代创建(高阶 Skill)
|
|
2
9
|
|
|
3
10
|
> **你负责**: 接收描述 → 智能命名 → 检查平台映射 → 创建迭代 → 推荐下一步。
|
|
@@ -1,3 +1,10 @@
|
|
|
1
|
+
---
|
|
2
|
+
name: spec-plan
|
|
3
|
+
description: >
|
|
4
|
+
执行计划生成。根据任务列表生成多阶段计划(Mermaid 图+风险评估+里程碑)。
|
|
5
|
+
Use when user says "规划" "计划" "plan" "排程".
|
|
6
|
+
allowed-tools: ["Bash", "Read", "Write", "Edit"]
|
|
7
|
+
---
|
|
1
8
|
# SpecCore Plan — 执行计划生成器
|
|
2
9
|
|
|
3
10
|
> **你负责**: 读取 Task 列表 → 你自己排程 → CLI 写入 plan.json。
|
|
@@ -1,3 +1,10 @@
|
|
|
1
|
+
---
|
|
2
|
+
name: spec-spec2doc
|
|
3
|
+
description: >
|
|
4
|
+
文档导出。Spec Markdown → Word/PDF/HTML/PPTX。
|
|
5
|
+
Use when user says "导出" "生成文档" "生成报告" "导出Word" "导出PDF".
|
|
6
|
+
allowed-tools: ["Bash", "Read", "Write", "Edit"]
|
|
7
|
+
---
|
|
1
8
|
# SpecCore spec2doc — AI 排版 + Pandoc 导出
|
|
2
9
|
|
|
3
10
|
> **你负责**: 审计 Spec 文档 → 你自己编排排版 → Pandoc 机械导出。
|
|
@@ -1,6 +1,14 @@
|
|
|
1
|
-
|
|
2
|
-
|
|
3
|
-
>
|
|
1
|
+
---
|
|
2
|
+
name: spec-task-create
|
|
3
|
+
description: >
|
|
4
|
+
交互式任务创建。10 种类型选择器 + AI 智能命名 + 内容澄清。
|
|
5
|
+
Use when user says "创建任务" "新建任务" "bug任务" "测试任务" "审查任务" "文档任务".
|
|
6
|
+
allowed-tools: ["Bash", "Read", "Write", "Edit"]
|
|
7
|
+
---
|
|
8
|
+
# SpecCore Task Create — 交互式任务创建
|
|
9
|
+
|
|
10
|
+
> **你负责**: 接收任意输入 → 内容澄清 → 智能命名 → 用 `execute_command` 创建任务 → 推荐下一步。
|
|
11
|
+
> **强制**: 用 `execute_command("speccore task new -n ...")` 执行,**绝不要把命令输出给用户复制**。
|
|
4
12
|
> **分支规则**: 创建任务后,Git 分支名 = `feature/Task-{id}-{任务名}`,与任务目录名完全一致。
|
|
5
13
|
|
|
6
14
|
## 任务类型选择器
|
|
@@ -148,12 +156,13 @@ Bug 类:
|
|
|
148
156
|
```
|
|
149
157
|
1. 接收输入 → 判断长度 → <80字必须澄清
|
|
150
158
|
|
|
151
|
-
2.
|
|
152
|
-
|
|
159
|
+
2. 语音命名 → 展示确认:
|
|
160
|
+
"📋 将创建: Task-003-login-feature,是否正确?[是/修改]"
|
|
161
|
+
|
|
162
|
+
3. execute_command("speccore task new --name '{name}' -t {type}")
|
|
153
163
|
|
|
154
|
-
|
|
155
|
-
speccore analyze --prompt -I {iter} --task {taskId}
|
|
156
|
-
speccore execute --prompt -t {taskId}
|
|
164
|
+
4. Write REQUIREMENT.md → 推荐下一步:
|
|
165
|
+
execute_command("speccore analyze --prompt -I {iter} --task {taskId}")
|
|
157
166
|
```
|
|
158
167
|
|
|
159
168
|
## 退出码
|
|
@@ -1,3 +1,11 @@
|
|
|
1
|
+
---
|
|
2
|
+
name: speccore-router
|
|
3
|
+
description: >
|
|
4
|
+
智能意图路由器。从自然语言识别意图 → 匹配 Skill → execute_command 执行。
|
|
5
|
+
核心原则:必须直接执行,决不输出命令文本。唯一下例外:"只看不跑"。
|
|
6
|
+
Use when user input is ambiguous or doesn't match any specific Skill trigger.
|
|
7
|
+
allowed-tools: ["Bash", "Read", "Write", "Edit"]
|
|
8
|
+
---
|
|
1
9
|
# SpecCore Router — 智能意图路由器
|
|
2
10
|
|
|
3
11
|
> **定位**:从用户自然语言识别意图,提取参数,**用 execute_command 自动执行** CLI 命令。
|