MdSpliter

大型知识文档分块查阅工具。当需要查阅知识库时，先通过索引匹配对应分块，避免一次性加载整个文件。 适用于: knowledge/*.md 文档分块、专题知识库索引、增量更新。

Security Scan

The skill is a purely instructional “knowledge splitter” that describes reading indexed Markdown chunks from a local knowledge/ directory and does not request external credentials, installs, or network endpoints — its declared purpose aligns with its instructions.

Purpose & Capability

Name/description (split large Markdown docs into indexed chunks) matches the manifest and SKILL.md instructions; no unrelated binaries, env vars, or external services are requested.

Instruction Scope

SKILL.md instructs the agent to read INDEX.md and chunk files under knowledge/* and to use a 'read' tool supporting offset/limit. This is consistent with the purpose but does require filesystem read access to the knowledge/ directory — the skill does not declare explicit config paths, which is acceptable for an instruction-only skill but worth noting.

Install Mechanism

No install spec or code is present (instruction-only). Nothing is downloaded or written to disk by the skill itself.

Credentials

The skill requests no environment variables or credentials. Its need for filesystem access is reasonable given the stated purpose; there are no unrelated secrets requested.

Persistence & Privilege

always is false and the skill does not request persistent or elevated privileges or modify other skills/config. Autonomous invocation is allowed by platform default but not unusually privileged here.

Assessment

This skill is instruction-only and simply describes reading an INDEX.md and topic chunks from a local knowledge/ directory. Before installing or enabling it: ensure you don't store secrets or sensitive files under the knowledge/ paths the skill might read; confirm what the agent's 'read' tool can access and whether you want to grant it filesystem read access; review any existing INDEX.md and chunk filenames to make sure only intended documents are present. No network endpoints, credentials, or install steps are requested by the skill, so the primary risk is accidental exposure of documents the agent can read — limit filesystem scope if needed.

Knowledge Splitter

大型知识文档分块查阅工具。

使用场景

当知识库文件过大（如 >5KB）时，应先分块存储，通过索引按需加载。

工作流程

1. 分块规则

按主题/类型拆分知识文件：

knowledge/
├── evomap-chunks/
│   ├── INDEX.md          # 索引文件
│   ├── 01-intro.md       # 主题1
│   ├── 02-protocol.md    # 主题2
│   └── ...
└── evomap_full.md       # 原始完整文件（可选保留）

2. 索引文件格式

每个知识库目录应有 INDEX.md：

# {知识库名} 索引

| 文件 | 标签 | 适用场景 |
|------|------|----------|
| 01-intro.md | intro, overview | 什么是X |
| 02-protocol.md | protocol, api | 协议相关 |

## 使用方法

1. 根据查询主题匹配标签
2. 读取对应的分块文件
3. 如需更详细内容，再查阅完整文件

3. 查阅流程

当需要查询知识时:
  1. 定位知识库目录 (如 knowledge/evomap-chunks/)
  2. 读取 INDEX.md 获取索引
  3. 根据查询主题匹配文件
  4. 读取对应的分块文件
  5. 如分块内容不足，再读取完整文件

工具

使用 read 工具读取文件：

• read 可指定 offset 和 limit 进行部分读取
• 大文件优先用分块读取

示例

查询 EvoMap GDI 评分:

1. 读取 knowledge/evomap-chunks/INDEX.md
2. 匹配标签 "gdi" → 05-gdi.md
3. 读取 knowledge/evomap-chunks/05-gdi.md

注意事项

• 分块文件保持 <2KB 为宜
• 索引文件包含所有分块的标签和用途
• 更新知识时同步更新索引
• 原始完整文件可选保留用于全文检索