Knowledge Graph & Relationships

Capture how decisions relate to patterns, how tasks track decisions, and how features depend on each other. Build a queryable relationship network that gives your AI real project topology.

Creating links

You: "Link the rate limiting pattern to Decision #14.
      The pattern implements the decision."

AI:  Relationship created:
     Pattern #8 (Rate limiting) implements Decision #14 (API protection)

     Now when you review Decision #14, you'll see which patterns
     implement it. When you work on the rate limiting pattern, you'll
     see the decision it addresses.

Exploring relationships

You: "Show me everything related to the authentication decision"

AI:  Linked items for Decision #7 (JWT authentication):

     Implements:
     • Pattern #3: Token validation middleware
     • Pattern #5: Token refresh strategy

     Tracks:
     • Progress #12: Implement JWT endpoints
     • Progress #18: Add token refresh logic

     Related Decisions:
     • Decision #12: Token expiry (7 days)
     • Decision #18: CORS policy

     Code Bindings:
     • src/auth/login.py
     • src/auth/middleware.py

Relationship types

Type	Meaning
`implements`	A pattern or task implements a decision
`related_to`	General association
`tracks`	Progress item tracks a decision
`blocks`	One item blocks progress on another
`clarifies`	Custom data clarifies a decision
`depends_on`	Dependency relationship
`supersedes`	A new decision supersedes an old one
`resolves`	A progress item resolves a decision
`derived_from`	An item was derived from another

MCP tools

link_engrams_items — create a relationship between two items
get_linked_items — retrieve all items linked to a given item (with optional filters)