
Hex Research
Stores and queries a local SQLite graph of research artifacts: hypotheses, goals, experimental runs, source quality ratings, audits, and generated maps. Part of the hex-skills ecosystem alongside hex-line, hex-graph, and hex-ssh. Writes its index to `.hex-skills/researchgraph/` and exposes 15 tools for creating, linking, and traversing research nodes. The agent can track which hypotheses led to which runs, score source reliability, and surface audit trails without hunting through scattered markdown files. Pairs naturally with the agile-workflow and codebase-audit-suite plugins when you need structured evidence for architectural decisions or experiment history. If you've ever tried to remember why you tested five different approaches last month, this keeps that context queryable instead of lost in chat logs.
- transport
- STDIO
- last updated
- May 8, 2026
Sponsors
Readme
Claude-Codex Skills
Open-source standalone agent skills, distributed as plugins for Claude Code and OpenAI Codex, help turn product ideas, code fixes, architecture reviews and deployments into scoped work with a verifiable result. Each workflow defines what to check, what it may change, and how to report evidence and remaining gaps.
Project page · Install · Full catalog
Choose a route
Choose by the result you need. Each skill works independently.
Details
- Repository
- github.com/levnikolaevich/claude-code-skills
- Package
@levnikolaevich/hex-research-mcp- Transport
- STDIO
- Registry
- MCP registry, active