Skip to content

Context Engine

A purpose-built server-side knowledge store that combines RAG and graph traversal behind an intent-based API. The Platform Framework never calls collections directly — it asks for what it needs.


What the Context Engine is not

  • Not a generic vector store
  • Not a graph database exposed directly
  • Not a key-value cache

It is an abstraction layer that combines RAG search, graph traversal, and tenant isolation internally, and exposes clean intent-based operations to callers.

Caller:          /resolve/skill — "give me the skill for this prompt"
Context Engine:  1. RAG search over Skill collection (semantic match)
                 2. Graph traverse: skill → context builders → guardrails → templates → tools
                 3. Merge, enrich, tenant-filter
                 4. Return one enriched response
Caller sees:     skill manifest + all dependencies resolved

Collections

Each collection stores a distinct knowledge type. Collections are never merged or blended — the CE decides which to query per API call.

CollectionWhat it storesWritten byRead by
Agent RegistryAgent manifests, capabilities (dense vectors)ADK/resolve/agent
Skill collectionSkill manifests (dense vectors per skill)ADK/resolve/skill, Phase 2C
Context Builder collectionContext builder specsADK/resolve/context
Tool collectionTool metadata, tool classesADK/resolve/tools
Guardrail collectionSafety rules per skill/domain/globalADKFormula execution
Template collectionOutput rendering templatesADKUI Agent
Domain Lens collectionExpert domain knowledge documentsADKAdvanced tier only
Domain Expert GraphControl relationship graphsADKAdvanced tier only
Playbook collectionPlaybook manifestsADK/resolve/playbook

Tenancy model

Tenant collection:   tenant_id = "tenant-xyz"    (customer-specific assets)
Global collection:   tenant_id = null             (platform-wide assets)

Resolution order:    tenant → global fallback

Guardrail exception: Both global and tenant guardrails always apply. Tenant guardrails add to global — they never replace them.


What the Context Engine does NOT own

ConcernOwnerReason
Estate dataContext Manager (client)Never leaves the customer's machine
Conversation historyContext Manager (client)Session state is client-side
CredentialsOS Keychain (client)Never enter the plugin layer
Artifacts and EvidenceContext Manager (client)Persisted locally in serial DB
LLM API callsDomain agentsCE builds context, agents call the LLM

Full API reference

See Context Engine API for all endpoints with request/response schemas.


Next steps

Escher — Agentic CloudOps by Tessell