Give coding agents persistent project memory with AgentMemory
Use AgentMemory to run a persistent memory service that lets coding agents recall project context, prior decisions, and reusable session knowledge.
npx skills add agentskillexchange/skills --skill give-coding-agents-persistent-project-memory-with-agentmemory
Use AgentMemory when a coding agent needs durable project memory across sessions instead of relying on a single chat transcript or ad hoc notes. The operator installs the AgentMemory service, starts the local memory server, connects a supported agent client such as Claude Code, Codex, Cursor, Gemini CLI, OpenClaw, or another MCP-capable tool, then lets the agent store and recall decisions, codebase facts, session summaries, and reusable context. Inputs are project activity, agent session notes, recall requests, and optional memory configuration; outputs are retrieved memories, persisted project context, searchable knowledge, and recall evidence that can guide later coding work. The scope boundary is persistent memory for coding-agent workflows, not a general database, note-taking product, vector-store listing, or model framework.
What this skill actually does
Inputs and prerequisites: Node.js, npm or npx, AgentMemory service, and a supported coding agent client such as Claude Code, Codex, Cursor, Gemini CLI, OpenClaw, or another MCP-compatible client.
Setup notes: Install globally with npm install -g @agentmemory/agentmemory, start the service with agentmemory, optionally seed the demo with agentmemory demo, then connect an agent with agentmemory connect claude-code or the equivalent supported client wiring. For one-off use, run npx -y @agentmemory/agentmemory@latest.
Source and verification boundary: use https://agent-memory.dev as the canonical reference before running the workflow; keep commands, API calls, CLI usage, and generated outputs reviewable against that upstream source.
Framework fit: publish this as a MCP workflow only when the operator can invoke the documented toolchain directly, rather than treating the upstream project as a generic product listing.