Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
What This Skill Does
ResearchVault is a local SQLite-backed state manager for agentic research workflows. It tracks artifacts, findings, and links across sessions, and supports branching hypotheses, automated synthesis via embeddings, and self-correcting verification missions.
Persistent local state with built-in synthesis and verification means agents can resume, cross-reference, and self-correct without external orchestration services.
When to use it
- Tracking findings across a multi-day research project
- Running parallel hypothesis branches to explore competing theories
- Ingesting Reddit threads and URLs into a structured knowledge base
- Automatically linking related findings using local embeddings
- Flagging low-confidence data for agent-driven verification
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: Research rising underground metal bands in Finland
- 1Initialize a new vault project with an objective via vault.py init
- 2Ingest Reddit and web sources using the scuttle command
- 3Run synthesize to discover links between findings using local embeddings
- 4Run verify plan to identify low-confidence findings and queue verification missions
- 5Start the MCP server to expose the vault to other agents or tools
A structured, interlinked knowledge base of findings with confidence scores and verified claims, accessible via MCP
Requirements
Accounts, API keys, or tools you or your AI assistant may need to set up while using this skill.