[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill-2737dd7e-5d5a-4e92-8121-0527de381f36":3,"$fwBxZRciHWiuTPU9lqW-CE13MVMJdUGa5X0axVQpVbLA":42},{"id":4,"title":5,"description":6,"categoryId":7,"moduleId":8,"tags":9,"prompt":10,"icon":11,"source":12,"sourceUrl":13,"authorId":14,"authorName":15,"isPublic":16,"stars":17,"runs":18,"createdAt":19,"updatedAt":19,"module":20,"category":27,"packages":33},"2737dd7e-5d5a-4e92-8121-0527de381f36","bdistill-behavioral-xray","X射线分析任何AI模型的行为模式——拒绝边界、幻觉倾向、推理风格、格式默认值。无需API密钥。","cat_coding_backend","mod_coding","sickn33,coding","---\nname: bdistill-behavioral-xray\ndescription: \"X-ray any AI model's behavioral patterns — refusal boundaries, hallucination tendencies, reasoning style, formatting defaults. No API key needed.\"\ncategory: ai-testing\nrisk: safe\nsource: community\ndate_added: \"2026-03-20\"\nauthor: FrancyJGLisboa\ntags: [ai, testing, behavioral-analysis, model-evaluation, red-team, compliance, mcp]\ntools: [claude, cursor, codex, copilot]\n---\n\n# Behavioral X-Ray\n\nSystematically probe an AI model's behavioral patterns and generate a visual report. The AI agent probes *itself* — no API key or external setup needed.\n\n## Overview\n\nbdistill's Behavioral X-Ray runs 30 carefully designed probe questions across 6 dimensions, auto-tags each response with behavioral metadata, and compiles results into a styled HTML report with radar charts and actionable insights.\n\nUse it to understand your model before building with it, compare models for task selection, or track behavioral drift over time.\n\n## When to Use This Skill\n\n- Use when you want to understand how your AI model actually behaves (not how it claims to)\n- Use when choosing between models for a specific task\n- Use when debugging unexpected refusals, hallucinations, or formatting issues\n- Use for compliance auditing — documenting model behavior at deployment boundaries\n- Use for red team assessments — systematic boundary mapping across safety dimensions\n\n## How It Works\n\n### Step 1: Install\n\n```bash\npip install bdistill\nclaude mcp add bdistill -- bdistill-mcp   # Claude Code\n```\n\nFor other tools, add bdistill-mcp as an MCP server in your project config.\n\n### Step 2: Run the probe\n\nIn Claude Code:\n```\n\u002Fxray                          # Full behavioral probe (30 questions)\n\u002Fxray --dimensions refusal     # Probe just one dimension\n\u002Fxray-report                   # Generate report from completed probe\n```\n\nIn any tool with MCP:\n```\n\"X-ray your behavioral patterns\"\n\"Test your refusal boundaries\"\n\"Generate a behavioral report\"\n```\n\n## Probe Dimensions\n\n| Dimension | What it measures |\n|-----------|-----------------|\n| **tool_use** | When does it call tools vs. answer from knowledge? |\n| **refusal** | Where does it draw safety boundaries? Does it over-refuse? |\n| **formatting** | Lists vs. prose? Code blocks? Length calibration? |\n| **reasoning** | Does it show chain-of-thought? Handle trick questions? |\n| **persona** | Identity, tone matching, composure under hostility |\n| **grounding** | Hallucination resistance, fabrication traps, knowledge limits |\n\n## Output\n\nA styled HTML report showing:\n- Refusal rate, hedge rate, chain-of-thought usage\n- Per-dimension breakdown with bar charts\n- Notable response examples with behavioral tags\n- Actionable insights (e.g., \"you already show CoT 85% of the time, no need to prompt for it\")\n\n## Best Practices\n\n- Answer probe questions honestly — the value is in authentic behavioral data\n- Run probes on the same model periodically to track behavioral drift\n- Compare reports across models to make informed selection decisions\n- Use adversarial knowledge extraction (`\u002Fdistill --adversarial`) alongside behavioral probes for complete model profiling\n\n## Related Skills\n\n- `@bdistill-knowledge-extraction` - Extract structured domain knowledge from any AI model\n\n## Limitations\n- Use this skill only when the task clearly matches the scope described above.\n- Do not treat the output as a substitute for environment-specific validation, testing, or expert review.\n- Stop and ask for clarification if required inputs, permissions, safety boundaries, or success criteria are missing.\n","","imported","https:\u002F\u002Fgithub.com\u002Fsickn33\u002Fantigravity-awesome-skills","user_system_seed","SkillOPIC",true,79,1492,"2026-05-16 13:08:31",{"id":8,"name":21,"slug":22,"icon":23,"description":24,"sort":25,"createdAt":26},"编程开发","coding","mdi-code-braces","代码生成、调试、审查，提升开发效率",2,"2026-05-16 12:53:40",{"id":7,"name":28,"slug":29,"icon":30,"description":31,"moduleId":8,"sort":25,"skillCount":32,"createdAt":26},"后端开发","backend","mdi-server","API、数据库、服务端架构",296,[34],{"id":35,"skillId":4,"version":36,"fileName":37,"fileSize":38,"filePath":39,"fileHash":40,"manifest":41,"createdAt":19},"72653f66-6357-4ff7-bfbe-167bcf878bb2","1.0.0","bdistill-behavioral-xray.zip",1802,"uploads\u002Fskills\u002F2737dd7e-5d5a-4e92-8121-0527de381f36\u002Fbdistill-behavioral-xray.zip","e15ae270c08ce6e2cdad4294599b7a36f00295a505d068b4befbce88c67d6c3b","[{\"path\":\"SKILL.md\",\"isDirectory\":false,\"size\":3588}]",{"code":43,"message":44,"data":45},200,"success",{"items":46,"stats":47,"page":50},[],{"averageRating":48,"totalRatings":48,"ratingCounts":49},0,[48,48,48,48,48],{"limit":51,"offset":48,"hasMore":52,"nextOffset":51,"ratedOnly":16},15,false]