
openanalyst
Turn your coding agent into a data analyst — DuckDB engine, auto-profiling, read-only SQL, and Vega-Lite charts. Ships as a DeepSeek Harness plugin (charts render live in the conversation) and an MCP server for Claude Code / Codex / Cursor. 把编码 Agent 变成数据分析师:dsh 插件对话内出图 + MCP server
What it doesAI
Turns coding agents into data analysts with DuckDB SQL, auto-profiling, and live Vega-Lite charts.
- DuckDB-powered read-only SQL engine
- Auto-profiling of datasets
- Renders Vega-Lite charts live in conversation
AI-generated from the repo README — for reference only.
Installation
dsh plugin --profile web add github:Chenmo0414/openanalystInstall method: GitHub · not yet tested in container (L3+)
Compatibility
| DSH Version | Status |
|---|---|
| not stated | Declared — not tested |
Requirements
- • Node.js: not stated
- • DSH: declared "not stated"
- • External credentials: none detected
Security Report
Automated static scan, not manual review.
No critical findings, but warning-level issues were detected (file writes, remote imports, base64 decoding, downloads).
0.1.0d3e2016860022026-08-20d3e201686002The scan result is valid for the scanned commit. New commits within 7 days are tolerated (the rating still counts); after 7 days without a rescan the badge shows outdated.
Disclaimer: automated static analysis, not a security guarantee. Always review what you install.
vet verdict: vet: clean · score 100
- R10 · info [heuristic] — 依赖清单:6 项(openanalyst, @deepseek-ai/dsh, @deepseek-ai/dsh-llm-mock-server, puppeteer-core, typescript, vitest,供 LLM 审计供应链;已知漏洞核对见后续 OSV 精确版本查询(网络失败静默降级))
- R3 · info [certain] — 只读 process 成员(能力触达面):process.env
- R3 · info [certain] (downgraded) — 只读 process 成员(能力触达面):process.exitCode
- R6 · info [heuristic] — 字符串特征:process 敏感成员引用
- R3 · info [certain] — 只读 process 成员(能力触达面):process.argv
vet is advisory and does not change dsh’s four-tier level.
• Static heuristic scan: done (15 files)
• Dependency vulnerabilities: requires deep audit (L3+)
• Permission sandboxing: requires runtime testing (L4+)
High-risk findings · 2 / 2
- warningHTTP request to a raw IP addressscripts/mock-llm-scripted.mjs:140console.log(`[mock] scripted LLM listening on http://127.0.0.1:${PORT}/v1 (${SCRIPT.length} entries)`)
- warningHTTP request to a raw IP addressscripts/shoot-conversation.mjs:25await page.goto('http://127.0.0.1:3080', { waitUntil: 'networkidle2', timeout: 60_000 })
Heuristic static scan — may produce false positives. Review the source yourself before trusting.
Activity
Last commit 2026-08-20 · activity: Active
• Repo created: 2026-08-20
• Stars: ★ 1 · Forks: 0
• Health: Active — committed within last 30 days