Memory
Search the OrchestKit knowledge graph, load session context, and view decision history via the memory MCP server while coding with Claude Code.
Overview
Memory is an agent skill most often used in Build (also Ship, Operate) that searches and visualizes OrchestKit knowledge-graph memory via MCP—pair with /ork:remember for writes.
Install
npx skills add https://github.com/yonatangross/orchestkit --skill memoryWhat is this skill?
- Subcommands: search, load, history, viz, status on unified read-side memory
- MCP tools: mcp__memory__search_nodes, mcp__memory__read_graph with Read/Grep/Glob/Bash
- Complements /ork:remember write-side; anti-triggers remember, save, store, commit
- Requires memory MCP server; Claude Code 2.1.148+; model hint haiku for low-effort reads
- 5 read subcommands: search, load, history, viz, status
- Claude Code 2.1.148+ required
Adoption & trust: 534 installs on skills.sh; 183 GitHub stars; 3/3 security scanners passed (skills.sh audits).
What problem does it solve?
You lose thread across sessions and cannot quickly search what the team or past-you decided without manually digging through chats and repos.
Who is it for?
Claude Code users with the memory MCP server who need search, load, history, or graph viz without invoking write-side remember flows.
Skip if: Persisting new facts or commits—use /ork:remember instead; also skip when anti-triggers apply (remember, save, store, commit, implement-only tasks without context need).
When should I use this skill?
Searching past decisions, loading context, visualizing the knowledge graph, or memory status—when user language matches recall, not remember/save/store.
What do I get? / Deliverables
You load searchable graph context, decision timelines, and optional Mermaid viz so the agent answers from stored memory before implementing new work.
- Search results from knowledge graph
- Loaded session context
- History or Mermaid graph visualization output
Recommended Skills
Journey fit
Spans multiple journey phases - primary shelf plus alternate fits below.
Canonical shelf is agent-tooling because the skill is defined around memory MCP read tools and .claude/memory paths during active development. Read-side memory (search, load, history, viz) is how you wire persistent context into the agent loop—not a one-off app feature.
Where it fits
Search_nodes for prior API shape choices before extending endpoints in the same repo.
Load session context from .claude/projects memory before wiring a new MCP enhancement.
View decision history to justify why a security tradeoff was accepted in an earlier review.
Recall monitoring or on-call decisions stored in the graph before changing alert thresholds.
How it compares
Read-side MCP skill paired with OrchestKit remember—not a generic notes app or filesystem-only grep.
Common Questions / FAQ
Who is memory for?
Solo builders on Claude Code 2.1.148+ with the memory MCP server who want OrchestKit read operations: search_nodes, read_graph, session load, and visualization.
When should I use memory?
In Build agent-tooling when grounding implementation; in Ship review when tracing past decisions; in Operate iterate when you need what we decided about errors, monitoring, or infra—triggered by search memory, knowledge graph, load context, or what did we decide.
Is memory safe to install?
It exposes read and limited Bash to project memory paths; check the Security Audits panel on this page and scope MCP access to repos you trust—memory may contain secrets or internal decisions.
SKILL.md
READMESKILL.md - Memory
# Memory - Read & Access Operations Unified read-side memory skill with subcommands for searching, loading, syncing, history, and visualization. > **Cross-session read strategy (Opus 4.7 / CC 2.1.111+):** Opus 4.7 reads filesystem memory more reliably than 4.6. When loading context at session start, prefer the layered read order: > 1. `~/.claude/projects/<slug>/memory/MEMORY.md` (durable index — load first, always) > 2. `.claude/chain/state.json` + most recent `NN-*.json` handoff (session continuation) > 3. MCP `mcp__memory__search_nodes` for anything the filesystem index doesn't answer (typed graph traversal) > > Layer 1 is cheap (small index file), Layer 2 is scoped (session-specific), Layer 3 is selective (only when needed). Avoid dumping the full knowledge graph into context — use the index to narrow the search first. ## Argument Resolution ```python SUBCOMMAND = "$ARGUMENTS[0]" # First token: search, load, history, viz, status QUERY = "$ARGUMENTS[1]" # Second token onward: search query or flags # $ARGUMENTS is the full string (CC 2.1.59 indexed access) ``` ## Usage ```bash /ork:memory search <query> # Search knowledge graph /ork:memory load # Load context at session start /ork:memory history # View decision timeline /ork:memory viz # Visualize knowledge graph /ork:memory status # Show memory system health ``` --- ## CRITICAL: Use AskUserQuestion When No Subcommand If invoked without a subcommand, ask the user what they want: ```python AskUserQuestion( questions=[{ "question": "What memory operation do you need?", "header": "Operation", "options": [ {"label": "search", "description": "Search decisions and patterns in knowledge graph", "preview": "```\nSearch Knowledge Graph\n──────────────────────\n query ──▶ mcp__memory ──▶ results\n\n Flags:\n --category Filter by type\n --agent Scope to agent\n --limit N Max results\n --global Cross-project\n```"}, {"label": "load", "description": "Load relevant context for this session", "preview": "```\nLoad Session Context\n────────────────────\n Auto-detect project ──▶\n ┌────────────────────┐\n │ Recent decisions │\n │ Active patterns │\n │ Project entities │\n └────────────────────┘\n Flags: --project, --global\n```"}, {"label": "history", "description": "Decision timeline + knowledge-graph viz (--mermaid)", "preview": "```\nDecision Timeline & Graph Viz\n─────────────────────────────\n ┌──── Feb 28 ────────────┐\n │ Used Postgres over Mongo│\n ├──── Feb 27 ────────────┤\n │ Adopted MVC pattern │\n └────────────────────────┘\n --mermaid ──▶ [Project]──uses─