[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-78267":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":10,"language":11,"languages":10,"totalLinesOfCode":10,"stars":12,"forks":13,"watchers":14,"openIssues":15,"contributorsCount":15,"subscribersCount":15,"size":15,"stars1d":15,"stars7d":15,"stars30d":16,"stars90d":15,"forks30d":15,"starsTrendScore":15,"compositeScore":17,"rankGlobal":10,"rankLanguage":10,"license":18,"archived":19,"fork":19,"defaultBranch":20,"hasWiki":21,"hasPages":19,"topics":22,"createdAt":10,"pushedAt":10,"updatedAt":31,"readmeContent":32,"aiSummary":33,"trendingCount":15,"starSnapshotCount":15,"syncStatus":34,"lastSyncTime":35,"discoverSource":36},78267,"crucible","Starlight143\u002Fcrucible","Starlight143","AI-native multi-agent research workflow: parallel evidence gathering, 7-direction debate, and risk-gated analysis — structured output, not one-shot prompts.","",null,"Python",110,12,9,0,84,3.34,"Other",false,"main",true,[23,24,25,26,27,28,29,30],"ai-workflow","autonomous-agents","debate","evidence-gathering","llm","multi-agent","openai","research-tool","2026-06-12 02:03:46","# Crucible\n\n\u003Cp align=\"center\">\n  \u003Cimg src=\"Crucible.png\" alt=\"Crucible\" width=\"640\">\n\u003C\u002Fp>\n\n\u003Cp align=\"center\">\n  \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Flicense-AGPL--3.0%20%2F%20Commercial-blue\" alt=\"License\">\n  \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fpython-3.10%2B-blue\" alt=\"Python\">\n  \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Ftests-3104%2B%20passing-brightgreen\" alt=\"Tests\">\n\u003C\u002Fp>\n\n\u003Cp align=\"center\">\n  \u003Cb>Languages:\u003C\u002Fb> \u003Cb>English (current)\u003C\u002Fb> · \u003Ca href=\"README_zh.md\">中文\u003C\u002Fa>\u003Cbr>\n  \u003Cb>Full manuals:\u003C\u002Fb> \u003Ca href=\"README_FULL.md\">Full English manual\u003C\u002Fa> · \u003Ca href=\"README_FULL_zh.md\">完整中文手冊\u003C\u002Fa>\n\u003C\u002Fp>\n\n---\n\n> **Turn \"I have an idea\" into \"runnable code + a 5-analyst risk review.\"**\n> An AI-native multi-agent research engine for investment research, SaaS product analysis, agent architecture evaluation, and academic paper reproduction.\n\n---\n\n## 30 seconds\n\n**Input** — a one-line idea, an existing project directory, or a paper title\n**Output** — a structured analysis report (consensus \u002F disagreement \u002F risks \u002F kill criteria) **plus runnable code**\n**How** — five specialist AI analysts + a quality gate that **halts the pipeline when evidence is too thin** (instead of confidently producing garbage)\n\n---\n\n## What will you use it for?\n\n| | |\n|---|---|\n| **🪙 Quant traders \u002F researchers**\u003Cbr>\"I have a strategy idea — is it worth pursuing?\"\u003Cbr>\u003Csub>→ Quant mode: research + backtest + tearsheet\u003C\u002Fsub> | **🛠️ SaaS \u002F product builders**\u003Cbr>\"I want a differentiated product — where are the real pain points?\"\u003Cbr>\u003Csub>→ SaaS mode: market + competitor + adoption barriers\u003C\u002Fsub> |\n| **📚 Academic researchers**\u003Cbr>\"I want to reproduce and ablate this paper.\"\u003Cbr>\u003Csub>→ Scientist mode: algorithm + reproducibility + benchmark\u003C\u002Fsub> | **🤖 AI agent developers**\u003Cbr>\"I'm building an agent — is the architecture sound?\"\u003Cbr>\u003Csub>→ Agent mode: state boundaries + replay safety\u003C\u002Fsub> |\n\n**Already have a project to fix?** Any of the four research modes can be paired with **Project Path mode** — point Crucible at a directory and it will analyse, fix bugs, and apply additive changes only (existing APIs preserved).\n\n→ Full mode reference: [README_FULL.md#pipeline-modes](README_FULL.md#pipeline-modes)\n\n---\n\n## Start in 3 steps\n\n### Recommended: the WebUI (no CLI flags to memorize)\n\n```bash\ngit clone https:\u002F\u002Fgithub.com\u002FStarlight143\u002Fcrucible.git\ncd crucible\npip install -r requirements.txt\n```\n\nThen double-click `launch_webui.bat` (Windows) — the browser opens automatically.\n\n**First-run setup:** open the Settings page and paste one API key:\n\n- [**OpenRouter**](https:\u002F\u002Fopenrouter.ai\u002F) — recommended; multi-model routing with USD cost tracking\n- [**Alibaba Coding Plan**](https:\u002F\u002Fhelp.aliyun.com\u002Fzh\u002Fmodel-studio\u002F) — token-only cost tracking\n- **Or run [Ollama](https:\u002F\u002Follama.ai) locally** — set `LLM_PROVIDER=ollama`; no key needed, zero ongoing cost\n\nThat's it. Pick a mode, type your idea (or paste a project path), hit Run.\n\n\u003Cdetails>\n\u003Csummary>\u003Cb>Prefer the CLI? Click to expand.\u003C\u002Fb>\u003C\u002Fsummary>\n\n```bash\n# Interactive mode\npython run_crucible.py\n\n# Dry-run: scan context without calling LLMs\npython run_crucible.py --dry-run\n\n# Self-check only\npython run_crucible.py --self-check\n\n# With Direction Debate\npython run_crucible.py --direction-debate\n\n# Full production scope with cost tracking\npython run_crucible.py --scope production --cost-trace --cost-report\n```\n\nFull flag reference: [README_FULL.md](README_FULL.md), or `python run_crucible.py --help`.\n\n\u003C\u002Fdetails>\n\n\u003Cdetails>\n\u003Csummary>\u003Cb>Production deployment (Gunicorn). Click to expand.\u003C\u002Fb>\u003C\u002Fsummary>\n\n```bash\npip install gunicorn\ngunicorn --config gunicorn_config.py \"webui.app:app\"\n```\n\nKey env overrides: `GUNICORN_BIND` (default `0.0.0.0:8080`), `GUNICORN_WORKERS`, `GUNICORN_TIMEOUT` (default `300`s — must exceed your longest pipeline run).\n\n\u003C\u002Fdetails>\n\n---\n\n## How it works\n\n```mermaid\ngraph LR\n    A[\"Your input\u003Cbr\u002F>idea \u002F path \u002F paper\"] --> B[\"Stage 0\u003Cbr\u002F>Web research\u003Cbr\u002F>+ citations\"]\n    B --> C[\"Stage 1\u003Cbr\u002F>3 parallel research lanes\u003Cbr\u002F>Market · Technical · Competitor\"]\n    C --> D{\"Stage 2\u003Cbr\u002F>Direction Debate\u003Cbr\u002F>(optional)\"}\n    D --> E[\"Stage 3\u003Cbr\u002F>5 specialist analysts\u003Cbr\u002F>+ Gate Controller\"]\n    E -->|insufficient evidence| F[\"⛔ Pipeline halts\"]\n    E -->|pass| G[\"Stage 4\u003Cbr\u002F>CodeGen + Quality Loop\"]\n    G --> H[\"📦 Report\u003Cbr\u002F>+ runnable code\"]\n```\n\nFive stages. Any stage can halt the pipeline if its quality gate fails — by design.\n\n→ Stage-by-stage breakdown: [README_FULL.md#stage-details](README_FULL.md#stage-details)\n\n---\n\n## Can I trust the output?\n\n- ✅ **Every claim is traceable to a cited source** — the Research Synthesizer drops unsupported claims to `unknowns` or flags them as `hallucination_flags`\n- ✅ **Every decision is traceable to a specific analyst finding** — full evidence chain preserved in output artifacts\n- ✅ **The pipeline halts automatically when evidence is insufficient** — no confident-but-wrong outputs\n- ✅ **2 588+ automated tests, 100% passing** — covering injection, SSRF, redaction, numerical stability, and cross-process races\n- ✅ **Backtest data integrity is enforced by default** — synthetic fallback data is rejected unless you explicitly opt in (`BACKTEST_REQUIRE_REAL_DATA=1` is the default)\n- ✅ **Pydantic-validated outputs at every stage** — downstream stages never parse free text\n\n---\n\n## License\n\n- **Free for personal, open-source, and academic use** — [AGPL v3](LICENSE)\n- **Commercial use** — closed-source distribution, proprietary SaaS that does not publish source, embedded use, or any deployment that cannot satisfy AGPL-3.0 requires a commercial license\n\nCommercial inquiries: **supervenus928@gmail.com** · See [COMMERCIAL_LICENSE.md](COMMERCIAL_LICENSE.md)\n\n---\n\n## More\n\n- 📘 [Full English manual](README_FULL.md) · [完整中文手冊](README_FULL_zh.md)\n- 🏗️ [Architecture](ARCHITECTURE.md)\n- 📝 [CHANGELOG](CHANGELOG.md)\n- 🤝 [Contributing](CONTRIBUTING.md)\n- ⚙️ [Configuration reference](.env.example)\n\nSee [CHANGELOG.md](CHANGELOG.md) for the full version history.\n","Crucible 是一个面向AI的多代理研究工作流工具，支持并行证据收集、七方向辩论以及风险控制分析，并以结构化输出而非一次性提示的形式提供结果。其核心功能包括通过五个专业AI分析师进行深入分析，并在证据不足时自动停止流程，避免产生无意义的结果。技术上，该项目使用Python编写，支持多种语言模型和API服务。Crucible适用于量化交易者\u002F研究人员、SaaS产品构建者、学术研究人员及AI代理开发者等场景，帮助他们从初步想法到可运行代码加五分析师的风险评估全过程。无论是投资策略验证、产品差异化分析、论文复现还是代理架构评估，Crucible都能提供强大的支持。",2,"2026-06-11 03:56:41","CREATED_QUERY"]