HomenShum/NodeBenchAI

Entity intelligence for any company, market, or question — not a chatbot that answers once, but a system that synthesizes with sources, turns each run into a reusable artifact, and watches for change later. Five-surface app + deep-work Workspace, Convex state, and a hosted public research MCP (npx nodebench-mcp).

Latest release v2.32.0 — v2.32.0 — Mobile UX + 4-Layer Grounding + Search Quality 92.5% · 28 Apr 2026

27 files for Claude Code, Codex and OpenCode: deep-diligence, deploy-and-launch, dogfood-loop, nodebench-benchmark-runner and 23 more — 35,037 tokens loaded in every session.

Open on GitHub nodebenchai.com

Instructions

NodeBenchAI AGENTS.md AGENTS.md C 30,304 tok
NodeBenchAI CLAUDE.md CLAUDE.md A 4,733 tok

Skills

demo-video-quality-loop .claude/skills/demo-video-quality-loop/SKILL.md A —
dep-convex-agent .claude/skills/dep-convex-agent/SKILL.md A 72 tok
dep-skills-howto .claude/skills/dep-skills-howto/SKILL.md A 50 tok
dep-tiptap-pm .claude/skills/dep-tiptap-pm/SKILL.md A 81 tok
dep-vega .claude/skills/dep-vega/SKILL.md A 80 tok
dep-xlsx .claude/skills/dep-xlsx/SKILL.md A 95 tok
live-qa-dogfood-loop .claude/skills/live-qa-dogfood-loop/SKILL.md A 17 tok
pipeline-operational-loop .claude/skills/pipeline-operational-loop/SKILL.md A 99 tok
proof-looping .claude/skills/proof-looping/SKILL.md A 118 tok

Commands

agentic-reliability-audit .claude/commands/agentic-reliability-audit.md A —
nodebench-qa .claude/commands/nodebench-qa.md A —
orchestrating-swarms .claude/commands/orchestrating-swarms.md B 59 tok
render-demo-video .claude/commands/render-demo-video.md A —
scenario-testing .claude/commands/scenario-testing.md A —
usage-report .claude/commands/usage-report.md A —

Agents

deep-diligence .claude/agents/deep-diligence.md A 49 tok
deploy-and-launch .claude/agents/deploy-and-launch.md A 41 tok
dogfood-loop .claude/agents/dogfood-loop.md A 25 tok
nodebench-benchmark-runner .claude/agents/nodebench-benchmark-runner.md A 26 tok
nodebench-changelog-updater .claude/agents/nodebench-changelog-updater.md A 24 tok
nodebench-deep-sim-analyst .claude/agents/nodebench-deep-sim-analyst.md A 40 tok
nodebench-frontend-polisher .claude/agents/nodebench-frontend-polisher.md A 38 tok
nodebench-supervisor .claude/agents/nodebench-supervisor.md A 31 tok
qa-dogfood .claude/agents/qa-dogfood.md A 44 tok
universal-product-diligence .claude/agents/universal-product-diligence.md A 43 tok

These files are HomenShum/NodeBenchAI's own configuration — they tell Claude Code, Codex and OpenCode how to work on this repository, so they are not mods to install elsewhere. Copy one as a starting point and replace the parts that are about this project.