[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-77672":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":13,"contributorsCount":13,"subscribersCount":13,"size":13,"stars1d":15,"stars7d":16,"stars30d":17,"stars90d":13,"forks30d":13,"starsTrendScore":18,"compositeScore":19,"rankGlobal":10,"rankLanguage":10,"license":10,"archived":20,"fork":20,"defaultBranch":21,"hasWiki":20,"hasPages":20,"topics":22,"createdAt":10,"pushedAt":10,"updatedAt":27,"readmeContent":28,"aiSummary":29,"trendingCount":13,"starSnapshotCount":13,"syncStatus":30,"lastSyncTime":31,"discoverSource":32},77672,"promptbeat","tophant-ai\u002Fpromptbeat","tophant-ai","Break your AI before they do.","",null,"MDX",676,0,125,80,143,542,240,4,false,"main",[23,24,25,26],"agents","ai-security","ai-security-testing","aisecurity","2026-06-12 02:03:44","# promptbeat\n\nRed-teaming toolkit for LLM safety evaluation. Generate adversarial test cases from curated seeds and evaluate model robustness.\n\n## Features\n\n- Seed-driven red-team test generation via promptfoo backend\n- Multi-provider evaluation with side-by-side comparison\n- Attack Success Rate metrics by risk category and provider\n- HTML report generation with interactive filtering\n- Self-contained release package with Python, Node.js, and promptfoo bundled\n\n## Installation\n\nDownload the release tarball for your platform, extract it, and run:\n\n```bash\ntar xf promptbeat-0.1.0-\u003Cplatform>.tar.gz\n.\u002Fpromptbeat-0.1.0-\u003Cplatform>\u002Fbin\u002Fpromptbeat --version\n```\n\nAvailable platforms:\n\n| File | Platform |\n| --- | --- |\n| `promptbeat-0.1.0-darwin-arm64.tar.gz` | macOS Apple Silicon |\n| `promptbeat-0.1.0-darwin-x64.tar.gz` | macOS Intel |\n| `promptbeat-0.1.0-linux-x64.tar.gz` | Linux x86_64 |\n\n## Quick Start\n\nCopy the sample configuration from `examples\u002Fbootstrap`, then run:\n\n```bash\nexport OPENAI_API_KEY=\"sk-...\"\n\n.\u002Fbin\u002Fpromptbeat generate \\\n  --target examples\u002Fbootstrap\u002Ftarget.yaml \\\n  --scenarios examples\u002Fbootstrap\u002Fscenarios.yaml \\\n  --seed-file examples\u002Fbootstrap\u002Fseeds.yaml \\\n  --provider openai:gpt-4o \\\n  --count 5\n\n.\u002Fbin\u002Fpromptbeat eval \\\n  --config artifacts\u002Fgenerate\u002Fgenerated_redteam.yaml\n\n.\u002Fbin\u002Fpromptbeat report \\\n  --eval-result artifacts\u002Feval\u002Fgenerated_redteam\u002Fevaluation_result.json\n```\n\n## Example Files\n\n- `examples\u002Fbootstrap\u002Ftarget.yaml`: target agent profile and safety boundaries\n- `examples\u002Fbootstrap\u002Fscenarios.yaml`: risk scenarios and judge configuration\n- `examples\u002Fbootstrap\u002Fseeds.yaml`: adversarial seed templates\n- `examples\u002Fbootstrap\u002Fproviders.yaml`: provider examples\n- `examples\u002Fbootstrap\u002Fpromptbeat.yaml`: end-to-end promptbeat configuration\n- `examples\u002Fbootstrap\u002Fpromptfoo.redteam.yaml`: promptfoo-compatible sample config\n\n## Requirements\n\n- An LLM provider API key such as `OPENAI_API_KEY`\n- No local Python, Node.js, or promptfoo installation is required when using the release package\n\n## Bundled Components\n\n| Component | Version |\n| --- | --- |\n| Python | 3.12.13 |\n| Node.js | 22.22.2 |\n| promptfoo | 0.121.9 |\n","promptbeat 是一个用于大型语言模型安全评估的红队工具包，旨在通过生成对抗性测试案例来检验模型的鲁棒性。其核心功能包括基于种子驱动的红队测试生成、多供应商并行评估与对比、按风险类别和供应商统计攻击成功率、以及生成带有交互式过滤功能的HTML报告。此外，该项目提供了一个自包含的发布包，集成了Python、Node.js和promptfoo，简化了安装与使用流程。适合需要对AI系统进行安全性检测的场景，尤其是那些关注于防止恶意输入导致模型行为异常的应用场合。",2,"2026-06-11 03:55:44","CREATED_QUERY"]