Openviking Memory

This gives Claude direct access to OpenViking's code repository search system through terminal commands. It's built for querying indexed codebases using semantic search, grep, or file globbing, all scoped under a viking:// URI namespace. The skill is densest when you're working with external libraries or frameworks you've pre-indexed, letting you ask conceptual questions like "how does auth work" instead of manually grepping. Scope matters here: narrowing searches to specific subdirectories dramatically improves speed and relevance. It also handles repo management (adding, removing, checking index status), though indexing runs async and can take 20+ minutes for large codebases. Most useful when you're integrating unfamiliar dependencies and need to understand implementation details without leaving the conversation.

skills.sh installs: 531
GitHub stars: 41.9K
last updated: Aug 21, 2026

View on GitHub

Install

$ npx -y skills add volcengine/openviking --skill openviking-memory --agent claude-code

Installs into .claude/skills of the current project.

Files

OpenViking Memory

OpenViking is a long-term semantic memory store addressed by viking:// URIs. This client has no lifecycle hooks, so nothing is recalled or captured automatically — you drive both halves of the loop with the openviking MCP tools.

Core tools, available on every supported deployment:

Recall: find, search, read, list, grep, glob
Persist: remember, add_resource
Maintain: forget, health

Some deployments register more than the core set — tree, write, edit, list_watches, cancel_watch. These are optional: which ones exist depends on the server version and hosting mode (the managed cloud service trims some of them). Check the session's registered tool list; if any optional tool is present, read references/optional-tools.md before using it. Never call a tool that is not registered, and do not fall back to raw HTTP. If no OpenViking tools are registered at all, continue without memory.