
{"id":9237,"date":"2026-06-17T19:54:44","date_gmt":"2026-06-17T11:54:44","guid":{"rendered":"https:\/\/infernews.com\/blog\/phoneharness-runtime-harness-for-mixed-action-phone-agents\/"},"modified":"2026-06-17T19:55:13","modified_gmt":"2026-06-17T11:55:13","slug":"phoneharness-runtime-harness-for-mixed-action-phone-agents","status":"publish","type":"post","link":"https:\/\/infernews.com\/blog\/phoneharness-runtime-harness-for-mixed-action-phone-agents\/","title":{"rendered":"PhoneHarness\uff1a\u91cd\u65b0\u91cf\u5ea6\u624b\u6a5f\u4ee3\u7406\u80fd\u529b"},"content":{"rendered":"\n<figure class=\"wp-block-image size-large\"><img decoding=\"async\" src=\"https:\/\/infernews.com\/blog\/wp-content\/uploads\/2026\/06\/phoneharness-demo-01-cli-status.gif\" alt=\"PhoneHarness CLI status demo\"\/><\/figure>\n\n\n\n<p class=\"wp-block-paragraph\">PhoneHarness \u662f\u4e00\u500b\u6df7\u5408\u52d5\u4f5c\u7684\u624b\u6a5f\u4ee3\u7406\u8a55\u6e2c\u6846\u67b6\u8207\u57fa\u6e96\uff0c\u89e3\u6c7a\u53ea\u9760 GUI \u9ede\u6309\u53bb\u8a55\u5206\u3001\u537b\u91cf\u4e0d\u5230\u771f\u5be6\u526f\u4f5c\u7528\u7684\u554f\u984c\u3002\u8ad6\u6587\u6307\u51fa\uff0c\u624b\u6a5f\u4efb\u52d9\u5f80\u5f80\u9700\u8981\u5728 GUI\u3001CLI \u8207 structured tools \u4e4b\u9593\u5207\u63db\uff0c\u55ae\u770b\u6700\u5f8c\u756b\u9762\u6703\u6f0f\u6389\u5f88\u591a\u95dc\u9375\u6b65\u9a5f\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u5b83\u7684\u505a\u6cd5\u662f\u628a\u884c\u52d5\u8def\u7531\u3001GUI \u59d4\u6d3e\u548c\u53ef\u8ffd\u8e64\u57f7\u884c\u8a18\u9304\u653e\u9032\u540c\u4e00\u500b\u6d41\u7a0b\u3002\u7576\u4efb\u52d9\u6709\u660e\u78ba\u53ef\u57f7\u884c\u8def\u5f91\u6642\uff0c\u7cfb\u7d71\u6703\u512a\u5148\u8d70 CLI \u6216 MCP \u5b8c\u6210\uff1b\u53ea\u6709\u5fc5\u8981\u6642\u624d\u4ea4\u7531 GUI worker \u900f\u904e\u622a\u5716\u4e92\u52d5\uff0c\u4ee4\u8a55\u6e2c\u66f4\u8cbc\u8fd1\u771f\u6b63\u624b\u6a5f\u5de5\u4f5c\u6d41\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u9019\u500b\u9805\u76ee\u7684\u65b0\u610f\u5728\u65bc\u628a\u300c\u80fd\u5426\u5b8c\u6210\u300d\u548c\u300c\u662f\u5426\u7559\u4e0b\u53ef\u9a57\u8b49\u8b49\u64da\u300d\u7d81\u5728\u4e00\u8d77\u3002JSONL traces \u548c HTML viewer \u4ee4\u5931\u6557\u53ef\u4ee5\u88ab\u62c6\u6210\u6a21\u578b\u63a8\u7406\u3001GUI \u5c0d\u9f4a\u3001\u74b0\u5883\u3001\u5de5\u5177\u6216 verifier \u4e0d\u4e00\u81f4\u5e7e\u985e\uff0c\u65b9\u4fbf\u627e\u51fa\u554f\u984c\u4f86\u6e90\uff0c\u800c\u4e0d\u662f\u53ea\u898b\u5230\u4e00\u500b\u5206\u6578\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u8ad6\u6587\u4e2d\u7684 PhoneHarness Bench \u5728 annotated evaluation split \u4e0a\u53d6\u5f97 75.0% pass rate\uff0c\u8f03\u6700\u5f37\u7684\u975e PhoneHarness \u8a2d\u5b9a\u9ad8 12.9 \u500b\u767e\u5206\u9ede\u3002\u9019\u8868\u793a\u5b83\u4e0d\u53ea\u662f\u6e2c\u8a66\u4ecb\u9762\u64cd\u4f5c\uff0c\u9084\u5728\u6e2c\u8a66\u4ee3\u7406\u600e\u6a23\u9078\u64c7\u52d5\u4f5c\u9762\uff0c\u5c0d\u505a\u624b\u6a5f\u81ea\u52d5\u5316\u3001\u88dd\u7f6e\u6e2c\u8a66\u3001\u6216\u9700\u8981\u53ef\u5be9\u8a08\u6d41\u7a0b\u7684\u5718\u968a\u90fd\u5e7e\u6709\u53c3\u8003\u50f9\u503c\u3002<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>\u6df7\u5408\u652f\u63f4 GUI\u3001CLI\u3001MCP tools\uff0c\u9069\u5408\u624b\u6a5f\u5de5\u4f5c\u6d41\u8a55\u6e2c<\/li>\n\n\n\n<li>\u512a\u5148 deterministic \u8def\u7531\uff0c\u6e1b\u5c11\u4e0d\u5fc5\u8981\u7684\u756b\u9762\u64cd\u4f5c<\/li>\n\n\n\n<li>\u53ef\u8ffd\u8e64 traces\uff0c\u65b9\u4fbf\u5b9a\u4f4d\u932f\u8aa4\u4f86\u6e90<\/li>\n\n\n\n<li>\u9069\u5408\u7814\u7a76 phone agents\u3001\u88dd\u7f6e\u81ea\u52d5\u5316\u8207\u5b89\u5168\u526f\u4f5c\u7528\u6aa2\u67e5<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">\u76f8\u95dc\u6a21\u578b\uff0f\u57fa\u6e96\u53ef\u4e00\u4f75\u7559\u610f\uff1aPhoneHarness\u3001PhoneHarness Bench\u3001AndroidWorld\u3001AppAgent\u3001Mobile-Agent-v2\u3001MobileAgentBench\u3001AndroidLab\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>GitHub\uff1a<\/strong> <a href=\"https:\/\/github.com\/PhoneHarness\/PhoneHarness\" rel=\"noopener noreferrer\">https:\/\/github.com\/PhoneHarness\/PhoneHarness<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>\u9805\u76ee\uff1a<\/strong> <a href=\"https:\/\/phoneharness.github.io\/\" rel=\"noopener noreferrer\">https:\/\/phoneharness.github.io\/<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"<p>\u5b83\u4e0d\u53ea\u770b\u4e0b\u4e00\u6b65\u9ede\u64ca\uff0c\u800c\u662f\u770b\u6709\u6c92\u6709\u771f\u6b63\u505a\u6210\u4e8b\u3002PhoneHarness \u628a GUI\u3001CLI \u548c\u5de5\u5177\u52d5\u4f5c\u653e\u5728\u540c\u4e00\u500b\u624b\u6a5f\u4ee3\u7406\u8a55\u6e2c\u6846\u67b6\u3002<\/p>\n","protected":false},"author":8,"featured_media":9236,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"ai_generated_summary":"","footnotes":""},"categories":[133,116,151,24,38,170,76,197],"tags":[],"class_list":["post-9237","post","type-post","status-publish","format-standard","hentry","category-133","category-agentic","category-mcp","category-aisoftware","category-38","category-170","category-76","category-framework"],"_links":{"self":[{"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/posts\/9237","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/users\/8"}],"replies":[{"embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/comments?post=9237"}],"version-history":[{"count":1,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/posts\/9237\/revisions"}],"predecessor-version":[{"id":9238,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/posts\/9237\/revisions\/9238"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/media\/9236"}],"wp:attachment":[{"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/media?parent=9237"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/categories?post=9237"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/tags?post=9237"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}