openviking-memory — Claude Code skill: what it does, how to install
Connects agents to OpenViking MCP long-term memory to retrieve relevant past knowledge at task start and save durable takeaways afterward.
npx skills add volcengine/OpenViking --skill openviking-memory -a claude-codeWhat openviking-memory does
This skill teaches the agent to manage long-term semantic memory in OpenViking via MCP tools. The client lacks lifecycle hooks. Because of this, nothing is saved or retrieved automatically. The model manually invokes find, search, and read against viking:// URIs before starting substantive tasks. Retrieved records remain advisory, so current workspace evidence always overrides old stored notes.
To persist durable context, the assistant calls remember with a concise summary or indexes external files with add_resource. It captures conventions, fixed bugs, and architectural choices, leaving out credentials and raw transcripts. The instructions strictly forbid fallback to raw HTTP: if OpenViking MCP tools are not registered in the session, the agent must simply work without memory. Older server deployments also reject the viking://~ alias with INVALID_URI, requiring explicit user paths instead.
What it helps with
- Retrieve previous bug fixes and procedures before starting complex tasks
- Store lasting conventions and system facts with
remember - Index external repositories and documentation files via
add_resource - Query project history by
viking://URIs without cluttering the context window
Example prompts for the agent
- Search OpenViking memory for our previous private registry deployment fix
- Remember that running integration tests requires a local redis instance
- Check viking://~/memories/experiences for prior CORS debugging steps
How to install openviking-memory
1Skills CLI (Vercel)
npx skills add volcengine/OpenViking --skill openviking-memory -a claude-code2Manually: put the files in place
~/.claude/skills/openviking-memory/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/openviking-memory/
1Skills CLI (Vercel)
npx skills add volcengine/OpenViking --skill openviking-memory -a codex2Manually: put the files in place
.agents/skills/openviking-memory/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add volcengine/OpenViking --skill openviking-memory -a cursor2Manually: put the files in place
.cursor/skills/openviking-memory/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install volcengine/OpenViking openviking-memory2Manually: put the files in place
.github/skills/openviking-memory/SKILL.md1Skills CLI (Vercel)
npx skills add volcengine/OpenViking --skill openviking-memory -a gemini-cli2Platform CLI
gemini skills install https://github.com/volcengine/OpenViking --path agent-plugins/skills/openviking-memoryManually: put the files in place
.devin/skills/openviking-memory/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add volcengine/OpenViking --skill openviking-memory -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
From the same set2
The whole set →Similar
Questions about openviking-memory
How to install openviking-memory in Claude Code?
Run npx skills add volcengine/OpenViking --skill openviking-memory -a claude-code.
What does the openviking-memory skill do?
It guides the agent on querying OpenViking memory with find and search, reading records from viking:// URIs, and persisting key lessons using remember.
Does openviking-memory capture context automatically?
No, the client has no lifecycle hooks. The assistant must trigger retrieval and persistence manually through registered OpenViking MCP tools.
