From memoryrelay
Use when building relationship maps between people, organizations, projects, or concepts that appear across multiple memories, or when recalling information that benefits from entity connections rather than flat search.
How this skill is triggered — by the user, by Claude, or both
Slash command
/memoryrelay:entity-and-contextThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Entities turn flat memory storage into a connected knowledge graph. Create entities for **named things that recur across multiple memories** and benefit from relationship tracking.
Entities turn flat memory storage into a connected knowledge graph. Create entities for named things that recur across multiple memories and benefit from relationship tracking.
| Tool | Signature | Purpose |
|---|---|---|
entity_create | entity_create(name, type, metadata?) | Create a new entity node |
entity_link | entity_link(entity_id, memory_id, relationship?) | Connect an entity to a memory |
entity_list | entity_list(limit?, offset?) | List entities with pagination |
entity_graph | entity_graph(entity_id, depth?, max_neighbors?) | Explore an entity's neighborhood |
memory_context | memory_context(query, token_budget?) | Enriched recall with entity connections (see memory-workflow skill) |
| Type | Examples |
|---|---|
person | Team members, stakeholders, external contacts |
place | Data centers, offices, deployment regions |
organization | Companies, departments, vendors |
project | Codebases, initiatives, products |
concept | Architecture patterns, business domains, protocols |
other | Anything that doesn't fit the above |
Create an entity when a named thing:
entity_create("AuthService", "concept", { "domain": "security" })memory_store(content, metadata) as usual (see memory-workflow skill)entity_link(entity_id, memory_id, "mentioned_in") to connect entity to each relevant memorymemory_context("authentication flow") returns memories enriched with linked entity dataentity_graph(entity_id, 2, 10) to see related memories and co-linked entities up to 2 hops awayUse memory_recall | Use memory_context |
|---|---|
| Simple keyword/semantic search | Need entity relationships in results |
| No entities involved | Entities are linked to relevant memories |
| Quick lookup of isolated facts | Understanding how things connect |
| Mistake | Fix |
|---|---|
| Creating entities for one-off mentions | Only create entities for things referenced across multiple memories |
| Not linking entities to memories | An unlinked entity is invisible to memory_context; always link after creating |
Using memory_recall when entities exist | Switch to memory_context for richer results that include entity connections |
| Creating duplicate entities | Call entity_list() first to check if the entity already exists |
| Linking everything to one entity | Keep links specific; an entity should connect only to directly relevant memories |
npx claudepluginhub memoryrelay/memory-relay-skills --plugin memoryrelayTraverses knowledge graph across memories, entities, and relationships for comprehensive context. Use before planning complex work, investigating concept connections, or answering 'what do you know about X'.
Use when user asks "what do you know about X", when planning complex work that spans multiple topics, when investigating how concepts connect across projects, or when simple memory queries don't provide enough context. Deep traversal of Forgetful MCP knowledge graph (mcp__forgetful__* tools).
Typed knowledge graph stored in sage-memory. Use when creating or querying structured entities (Person, Project, Task, Event, Document), linking related objects, checking dependencies, planning multi-step actions as graph transformations, or when skills need to share structured state. Trigger on "remember that X is Y", "what do I know about", "link X to Y", "show dependencies", "what blocks X", entity CRUD, cross-skill data access, or any request involving structured relationships between things. Also trigger when the memory skill is active and the agent needs typed structure beyond flat prose.