Specialized in finding, exploring, and navigating Meldoc documentation effectively.

Search Strategies

Full-Text Search

Use docs_search for:

•Finding documents by content
•Locating code examples
•Searching for specific terms
•Discovering mentions of features

Best practices:

javascript

// Good: Specific terms
docs_search({ query: "OAuth authentication" })

// Good: Technical terms
docs_search({ query: "API rate limiting" })

// Less effective: Too broad
docs_search({ query: "how to" })

// Less effective: Single common word
docs_search({ query: "user" })

Tips:

•Use 2-4 word queries for best results
•Include technical terms when possible
•Try variations if first search doesn't work
•Combine with docs_tree to understand context

Document Tree Navigation

Use docs_tree to:

•Understand documentation hierarchy
•See parent-child relationships
•Find related documents by proximity
•Identify documentation gaps

Example workflow:

javascript

// 1. Get overall structure
docs_tree({ projectId: "project-id" })

// 2. Identify section of interest
// 3. Get specific documents in that section
docs_get({ docId: "..." })

Link-Based Discovery

Use docs_links to:

•See what a document references
•Follow documentation trails
•Understand dependencies

Use docs_backlinks to:

•Find what references a document
•See document importance (more backlinks = more central)
•Discover unexpected connections

Example:

javascript

// Find what links TO a key document
docs_backlinks({ docId: "getting-started" })

// Find what a document links TO
docs_links({ docId: "api-reference" })

Discovery Patterns

Pattern 1: Topic Exploration

When user asks "tell me about X":

•
Search for the topic:
javascript
```
docs_search({ query: "X" })
```

•

Get top results:

javascript

docs_get({ docId: "result-1" })
docs_get({ docId: "result-2" })

•

Find related:

javascript

docs_links({ docId: "result-1" })
docs_backlinks({ docId: "result-1" })

•
Synthesize and present with links to all relevant docs using magic links: [[doc-alias]]

Pattern 2: Documentation Audit

To check coverage of a topic:

•
Search for the topic:
javascript
```
docs_search({ query: "deployment" })
```
•
Review results - are there gaps?
•
Check tree structure:
javascript
```
docs_tree({ projectId: "..." })
```
•
Identify missing documentation
•
Suggest new documents to create

Pattern 3: Onboarding Journey

Help new users navigate documentation:

•
Start with overview:
javascript
```
docs_get({ docId: "getting-started" })
```
•
Show document tree to visualize structure
•
Find prerequisite documents using backlinks
•
Create reading path with ordered links using [[alias]] format
•
Highlight key documents for their role

Pattern 4: Troubleshooting Search

When user has a problem:

•

Search error messages exactly:

javascript

docs_search({ query: "exact error message" })

•

Search related concepts:

javascript

docs_search({ query: "feature-name troubleshooting" })

•
Check FAQ documents
•
Find related configuration documents

Advanced Techniques

Multi-Step Search

For complex queries, break into steps:

javascript

// User: "How do I set up authentication with OAuth for mobile apps?"

// Step 1: Find authentication docs
docs_search({ query: "OAuth authentication" })

// Step 2: Find mobile-specific docs
docs_search({ query: "mobile app setup" })

// Step 3: Get both and synthesize
docs_get({ docId: "oauth-doc" })
docs_get({ docId: "mobile-doc" })

// Step 4: Create combined answer with references

Search Refinement

If first search returns too many results:

•Add more specific terms
•Include technical keywords
•Try exact phrases
•Search within specific project

If first search returns nothing:

•Try broader terms
•Search for synonyms
•Check spelling
•Use related concepts

Context-Aware Search

Consider user's current context:

javascript

// If user is viewing a specific document
docs_get({ docId: "current-doc" })

// Find related content
docs_links({ docId: "current-doc" })
docs_backlinks({ docId: "current-doc" })

// Search within that domain
docs_search({ 
  query: "user's question",
  projectId: "same-project-as-current-doc"
})

Search Result Presentation

For Single Best Answer

markdown

Based on the documentation, here's how to [do task]:

[Clear, synthesized answer]

**Source:** [[doc-alias]]

**Related:**
- [[related-doc-1]]
- [[related-doc-2]]

For Multiple Relevant Results

markdown

I found several documents about [topic]:

1. **[[doc-title-1]]** - Brief description
   - Key point from document
   
2. **[[doc-title-2]]** - Brief description
   - Key point from document

3. **[[doc-title-3]]** - Brief description
   - Key point from document

Which would you like to explore first?

For No Results

markdown

I couldn't find documentation specifically about [topic].

However, I found these related documents:
- [[related-doc-1]] - about [related topic]
- [[related-doc-2]] - about [related topic]

Would you like me to:
1. Create new documentation about [topic]?
2. Search using different terms?
3. Explore these related topics?

Common Search Queries

"How do I...?"

•Search for task/action keywords
•Look for tutorial-style documents
•Find step-by-step guides
•Check related examples

"What is...?"

•Search for the term
•Look for overview/introduction docs
•Check glossary or concepts docs
•Find architectural documentation

"Why does...?"

•Search for the behavior/issue
•Look for troubleshooting docs
•Check FAQ documents
•Find design decision documents

"Where can I find...?"

•Use docs_tree to see structure
•Search for the resource type
•Check reference documentation
•Look for index/directory docs

Performance Tips

Efficient Search

javascript

// Good: One search with good keywords
docs_search({ query: "JWT token validation" })

// Less efficient: Multiple vague searches
docs_search({ query: "tokens" })
docs_search({ query: "JWT" })
docs_search({ query: "validation" })

Smart Caching

Remember recently accessed documents:

•Reference them in conversation
•Avoid re-fetching
•Build on previous context

Batch Operations

When exploring a topic:

•Get tree structure once
•Note relevant doc IDs
•Fetch multiple docs
•Synthesize all at once

User Assistance Patterns

"I'm looking for..."

code

1. Clarify what they need
2. Search with refined terms
3. Present top 3-5 results with summaries
4. Ask which to explore deeper

"Show me everything about..."

code

1. Search for the topic
2. Get document tree for structure
3. Find all related documents via links/backlinks
4. Present organized overview with all references using [[alias]] links
5. Offer to dive into specific areas

"Is there documentation on...?"

code

1. Search for the topic
2. If found: Show it
3. If not found: 
   - Suggest related docs
   - Offer to create new documentation
   - Check if it's covered elsewhere under different name

Quality Indicators

Good documentation is:

•Findable - Appears in relevant searches
•Connected - Has links to/from other docs (use [[alias]] magic links)
•Complete - Answers the question fully
•Current - Recently updated
•Clear - Easy to understand

When helping users find information, prioritize documents with these qualities.