search-doc

name	search_doc
description	Full-text search across all repositories
category	search
version	1.1.0
author	DocSys Team
permissions	["read:document"]
tags	["search","find","query","full-text"]

Search Documents

Full-text search across all accessible repositories.

Triggers

Search, search docs, 搜索, 查找, 检索, find documents, find files, search for, 搜索文档, 查找文件, 查询, search everything, global search, find, 全文搜索, 文件搜索, 文件查找

Workflow

1. Parse query: Extract search keywords from user input. Multi-word queries: pass as space-separated arguments (e.g., docsys search machine learning papers). Phrase search: wrap in double quotes if supported (e.g., docsys search "project timeline").
2. Optional vid filter: If user specifies a repository name → call docsys repos list to resolve name to vid; pass vid to CLI.
3. Call CLI: docsys search <query> [vid]
4. Zero-results handling: If no results → suggest broadening: remove filters, try fewer keywords, or search without a vid filter. Example: "No results for 'complex phrase'. Try 'complex' alone or without repo filter."
5. Overflow handling: If too many results → return top 20 sorted by relevance, note "showing 20 of N results". Suggest adding a vid filter or more specific terms.
6. Format output: Show per result: repository name, filename, path, size, date.

CLI Command

docsys search <query> [vid]

Parameters

Parameter	Type	Required	Description
query	string	yes	Search keywords (space-separated for multi-word)
vid	number	no	Limit search to specific repository

Examples

Example 1

User: "search for project documents" Skill triggers → calls: docsys search project Output: List of matching documents with repository, path, size, date

Example 2

User: "搜索包含 Q4 报告的文档" Skill triggers → calls: docsys search Q4 报告 Output: 匹配的文档列表

Example 3

User: "find machine learning papers" Skill triggers → Step 1: query="machine learning" → calls: docsys search machine learning papers Output: Matching documents

Example 4

User: "搜索 Claude AI 相关文档" Skill triggers → Step 1: query="Claude AI" → calls: docsys search Claude AI Output: 匹配文档列表

Example 5 (zero results → refinement)

User: "搜索 项目计划 2024 完整版" Skill triggers → Step 1: query="项目计划 2024 完整版" → Step 3: 0 results → Step 4: suggests "No results. Try '项目计划' alone or broaden query."

Output Format

Field	Type	Description
repo	string	Repository name
filename	string	Document filename
path	string	File path within repo
size	string	File size
date	string	Last modified date
relevance	string	Relevance score (if available)

Success (EN): {repo} | {filename} | {path} | {size} | {date} Success (中文): {repo} | {filename} | {path} | {size} | {date} Zero results: ⚠️ No results for "{query}". Try broader terms or remove repo filter. Overflow: 📋 Showing top 20 of {N} results. Add vid filter or more specific terms to narrow. Error: "Query is required" | "Search service unavailable"

Error Handling

Error	Cause
Query is required	Empty search query
Repository not found	Invalid vid
Search service unavailable	DocSystem search is down
Too many results	Narrow query

See references/ for related skills and API docs.