knowledgine-suggest

name	knowledgine-suggest
description	Get contextual knowledge suggestions based on the current work context — the file being edited, the task being performed, or the error being investigated. Invoke at the start of a work session, when opening a new file, or when beginning a feature. Surfaces relevant past knowledge proactively without requiring a specific query.

knowledgine-suggest

Purpose

Surface relevant knowledge from the knowledge base based on contextual signals extracted from the current work, rather than a specific search query. Suggest combines search and graph traversal to bring forward knowledge you might not know to search for.

When to Use

• Starting a session — Get context about the area you will work in
• Opening a new file — Discover past knowledge about that component
• Beginning a feature — Find related patterns and past decisions
• Before making changes — Check for warnings or prior art on the topic
• Exploring an unfamiliar module — Find related entities and notes

When NOT to Use

• When you already have a specific query in mind (use knowledgine-recall instead)
• When working on trivial mechanical tasks with no design ambiguity
• After already calling suggest for the same file in the same session

How to Get Suggestions (MCP Tools)

Step 1: Search by context

search_knowledge(
  query: string,   // Derived from file path, task description, or component name
  mode: "hybrid",  // Hybrid gives best results for context-based queries
  limit: 10
)

Step 2: Find related notes (if you have a note ID or file path)

find_related(
  noteId?: string,    // ID of a relevant note found in step 1
  filePath?: string,  // Current file path being worked on
  limit?: number,     // Max results (default 10)
  maxHops?: number    // Graph traversal depth (default 2)
)

How to Get Suggestions (CLI Alternative)

knowledgine suggest "<context description>"
knowledgine suggest --file src/commands/setup.ts

Extracting Context

The quality of suggestions depends on the context you provide. Extract signals from:

Signal	Example
Current file path	src/commands/setup.ts
Component or class name	SetupCommand, KnowledgeRepository
Task description	"add TOML config support to setup command"
Feature area	"MCP configuration", "entity extraction"
Error message	"TypeError: Cannot set properties of undefined"

Step-by-Step Instructions

1. Extract context signals from the current task (see context-patterns.md)
2. Build a query — Combine the most specific signals into a search query
3. Call search_knowledge with mode: "hybrid" and the derived query
4. Review results — Read the top 3–5 most relevant notes
5. Traverse related notes — Call find_related with a relevant noteId or the current filePath
6. Summarize findings — Note any relevant warnings, patterns, or decisions found
7. Proceed informed — Use the discovered context to guide your implementation

Best Practices

• Run suggest once at session start for the primary area of work
• Combine with knowledgine-recall for targeted queries as specific issues arise
• Prioritize notes tagged with bug-fix or troubleshooting — they contain warnings
• Notes tagged design-decision are especially valuable before making changes

Reference Files

• See context-patterns.md for how to extract effective context from your current work