​

Dependency Chain

​

What it does

​

Parameters

​

Returns

• chain: Array of dependencies organized by depth level
• totalNodes: Total count of nodes in the dependency chain

• depth: Level in the dependency tree (1 = direct dependencies)
• label: Name of the dependency
• nodeType: Type of node (person, project, goal, tool, etc.)
• relationType: Type of relationship (depends_on, needs, etc.)
• confidence: Confidence score for the relationship
• evidenceMemoryIds: Memories that establish this dependency

{
  "entity": "Q2 launch",
  "chain": [
    {
      "depth": 1,
      "label": "API redesign",
      "nodeType": "project",
      "relationType": "depends_on",
      "confidence": 0.92,
      "evidenceMemoryIds": ["mem123"]
    },
    {
      "depth": 2,
      "label": "Database schema update",
      "nodeType": "project",
      "relationType": "depends_on",
      "confidence": 0.87,
      "evidenceMemoryIds": ["mem456"]
    }
  ],
  "totalNodes": 7
}

​

Examples

​

Plan project work

crystal_dependency_chain(entity="Q2 launch")

​

Understand blocking issues

crystal_dependency_chain(
  entity="mobile app release",
  maxDepth=2
)

​

Trace goal dependencies

crystal_dependency_chain(
  entity="improve API performance",
  maxDepth=5
)

​

When to use

• Plan work — Identify all tasks needed before starting a project
• Find critical path — Understand which dependencies are blocking progress
• Estimate timelines — See what work must be sequential vs. parallel
• Identify bottlenecks — Spot dependencies that many things rely on
• Risk assessment — Understand how changes to a dependency affect downstream work

​

Errors

​

Notes

​

Peer scoping (new in 0.7.15)

• Depth is clamped to 1-5 automatically (min: 1, max: 5)
• Default depth of 3 captures most practical dependency trees
• Results are organized by level for easy scanning
• Evidence memory IDs link back to where dependencies were recorded
• Total node count includes all nodes across all depths
• Circular dependencies are handled gracefully and will not cause infinite loops
• API-key clients use a recall-based fallback if the full graph query is unavailable