[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M698Y8QMCBMW9MZB0A7KF1T4":3,"skill-same:01M698Y8QMCBMW9MZB0A7KF1T4":100},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":28,"publishedAt":49,"linkOnly":50,"markdownContent":51,"markdownBody":52,"frontMatter":53,"refName":54,"commitSha":55,"skillDirectory":56,"forksCount":57,"repositoryHtmlUrl":58,"repositoryHomepageUrl":59,"ownerHtmlUrl":60,"ownerBlogUrl":61,"ownerDescription":62,"updatedAt":63,"install":64,"resources":77},"01M698Y8QMCBMW9MZB0A7KF1T4","adk-debug",null,"Diagnoses misbehaving ADK agents by inspecting sessions, events, tool calls, and the exact request that reached the model. Covers the `adk run` CLI and the `adk web` dev server with its session, trace, and debug HTTP endpoints. Use when an agent returns the wrong answer, ignores a tool or swallows a tool error, hangs, loops, emits raw JSON instead of calling tools, is not discovered by `adk web`, when a sub-agent cannot see the parent conversation, or when you need the LLM request\u002Fresponse, token counts, or logs for a run. Don't use for how ADK is designed internally (use `adk-architecture`), for building a new agent or workflow (use `adk-agent-builder`), for environment or dependency setup failures (use `adk-setup`), or for lint and style nits (use `adk-style`).","google\u002Fadk-python","google","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F1342004?v=4","01M698VYEFDM0VJBZDYYJ2T83S","Google","01M698XZC2HCATT8VD5VAVZDTP",22217,"Apache-2.0","https:\u002F\u002Fgithub.com\u002Fgoogle\u002Fadk-python\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fadk-debug\u002FSKILL.md",[18,23],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000023","Skill与智能体开发","SCENARIO","使用场景",{"categoryNo":24,"categoryName":25,"taxonomyType":26,"taxonomyTypeDesc":27},"01JBSEEDCAT000000000000030","开发工程师","PROFESSION","适用职业",[29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48],"adk debug","adk","debug","when","cli","request","you","loops","error","new agent","conversation","builder","dependency","model","how","server","logs","run","architecture","workflow","2026-11-10 10:12:30",false,"---\nname: adk-debug\ndescription: >-\n  Diagnoses misbehaving ADK agents by inspecting sessions, events, tool calls,\n  and the exact request that reached the model. Covers the `adk run` CLI and the\n  `adk web` dev server with its session, trace, and debug HTTP endpoints. Use\n  when an agent returns the wrong answer, ignores a tool or swallows a tool\n  error, hangs, loops, emits raw JSON instead of calling tools, is not\n  discovered by `adk web`, when a sub-agent cannot see the parent conversation,\n  or when you need the LLM request\u002Fresponse, token counts, or logs for a run.\n  Don't use for how ADK is designed internally (use `adk-architecture`), for\n  building a new agent or workflow (use `adk-agent-builder`), for environment or\n  dependency setup failures (use `adk-setup`), or for lint and style nits (use\n  `adk-style`).\n---\n\n# Debugging ADK agents\n\nTwo entry points. Default to `adk run`: one process, no server, and `--jsonl`\noutput that pipes straight into `grep` or `python3`. Switch to `adk web` when\nyou need the browser UI, a persisted session you can click through, or the\ntrace endpoints that expose the exact LLM request.\n\n## First moves\n\n1. Reproduce headlessly: `adk run --jsonl {agent_dir} \"{query}\"`. Without\n   `--jsonl`, `adk run` prints only text parts — tool calls and tool errors are\n   invisible.\n2. Read the log file. `adk run` writes to `\u002Ftmp\u002Fagents_log\u002Fagent.latest.log` and\n   nothing to the terminal; `adk web` does the opposite. See\n   [logs-and-traces.md](references\u002Flogs-and-traces.md).\n3. Match the symptom in [failure-modes.md](references\u002Ffailure-modes.md) before\n   reading source — most reports are one of a handful of known shapes.\n4. If the text is fine but the routing is not, dump the events and read\n   `author`, `branch`, `nodeInfo.path`, and `actions` —\n   [event-flow.md](references\u002Fevent-flow.md).\n5. If the model itself misbehaved, read what it actually received from the\n   `call_llm` span rather than guessing from the agent definition —\n   [logs-and-traces.md](references\u002Flogs-and-traces.md).\n\n## References\n\n- [cli-run.md](references\u002Fcli-run.md) — `adk run` flags, the JSONL event shape,\n  multi-turn and human-in-the-loop resume, exit codes, driving a `Runner` from\n  Python.\n- [web-api.md](references\u002Fweb-api.md) — starting `adk web`, listing and reading\n  sessions over HTTP, posting test messages to `\u002Frun_sse`.\n- [logs-and-traces.md](references\u002Flogs-and-traces.md) — log levels and where\n  each command writes them, the trace endpoints, span attributes, and the env\n  vars that control whether prompts appear in spans.\n- [failure-modes.md](references\u002Ffailure-modes.md) — ADK-specific symptoms with\n  the cause and a concrete check for each.\n- [event-flow.md](references\u002Fevent-flow.md) — how an invocation becomes events,\n  callback order, the event fields that matter, and where each stage lives in\n  the source.\n\n## Ground rules\n\n- Leave sessions in place when you finish. The user may still want to open them\n  in the web UI, and `adk web` has no undelete.\n- Delete any throwaway agent you created for a repro, unless the user asked to\n  keep it.\n- Reach for a unit test in `tests\u002Funittests\u002F` when the bug is inside one\n  component, and for a sample under `contributing\u002Fsamples\u002F` (see\n  `adk-sample-creator`) when it only reproduces with runner, agent, and workflow\n  wired together.\n","# Debugging ADK agents\n\nTwo entry points. Default to `adk run`: one process, no server, and `--jsonl`\noutput that pipes straight into `grep` or `python3`. Switch to `adk web` when\nyou need the browser UI, a persisted session you can click through, or the\ntrace endpoints that expose the exact LLM request.\n\n## First moves\n\n1. Reproduce headlessly: `adk run --jsonl {agent_dir} \"{query}\"`. Without\n   `--jsonl`, `adk run` prints only text parts — tool calls and tool errors are\n   invisible.\n2. Read the log file. `adk run` writes to `\u002Ftmp\u002Fagents_log\u002Fagent.latest.log` and\n   nothing to the terminal; `adk web` does the opposite. See\n   [logs-and-traces.md](references\u002Flogs-and-traces.md).\n3. Match the symptom in [failure-modes.md](references\u002Ffailure-modes.md) before\n   reading source — most reports are one of a handful of known shapes.\n4. If the text is fine but the routing is not, dump the events and read\n   `author`, `branch`, `nodeInfo.path`, and `actions` —\n   [event-flow.md](references\u002Fevent-flow.md).\n5. If the model itself misbehaved, read what it actually received from the\n   `call_llm` span rather than guessing from the agent definition —\n   [logs-and-traces.md](references\u002Flogs-and-traces.md).\n\n## References\n\n- [cli-run.md](references\u002Fcli-run.md) — `adk run` flags, the JSONL event shape,\n  multi-turn and human-in-the-loop resume, exit codes, driving a `Runner` from\n  Python.\n- [web-api.md](references\u002Fweb-api.md) — starting `adk web`, listing and reading\n  sessions over HTTP, posting test messages to `\u002Frun_sse`.\n- [logs-and-traces.md](references\u002Flogs-and-traces.md) — log levels and where\n  each command writes them, the trace endpoints, span attributes, and the env\n  vars that control whether prompts appear in spans.\n- [failure-modes.md](references\u002Ffailure-modes.md) — ADK-specific symptoms with\n  the cause and a concrete check for each.\n- [event-flow.md](references\u002Fevent-flow.md) — how an invocation becomes events,\n  callback order, the event fields that matter, and where each stage lives in\n  the source.\n\n## Ground rules\n\n- Leave sessions in place when you finish. The user may still want to open them\n  in the web UI, and `adk web` has no undelete.\n- Delete any throwaway agent you created for a repro, unless the user asked to\n  keep it.\n- Reach for a unit test in `tests\u002Funittests\u002F` when the bug is inside one\n  component, and for a sample under `contributing\u002Fsamples\u002F` (see\n  `adk-sample-creator`) when it only reproduces with runner, agent, and workflow\n  wired together.\n",{"name":5,"description":7},"main","8cd3b12c8a5d658c5e0d3ea0731d8be34a092a4a",".agents\u002Fskills\u002Fadk-debug",4356,"https:\u002F\u002Fgithub.com\u002Fgoogle\u002Fadk-python","https:\u002F\u002Fgoogle.github.io\u002Fadk-docs\u002F","https:\u002F\u002Fgithub.com\u002Fgoogle","https:\u002F\u002Fopensource.google\u002F","Google ❤️ Open Source","2026-11-14 14:38:36",{"source":8,"skillName":5,"commands":65,"prompt":75,"zipAvailable":76},[66,69,72],{"tool":67,"command":68},"npx","npx skills add google\u002Fadk-python --skill adk-debug",{"tool":70,"command":71},"bunx","bunx skills add google\u002Fadk-python --skill adk-debug",{"tool":73,"command":74},"pnpm","pnpm dlx skills add google\u002Fadk-python --skill adk-debug","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fgoogle\u002Fadk-python\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fadk-debug\u002FSKILL.md",true,[78,84,88,92,96],{"path":79,"resourceType":80,"resourceTypeDesc":81,"sizeBytes":82,"sourceHtmlUrl":83},"references\u002Fcli-run.md","REFERENCE","参考文档",5007,"https:\u002F\u002Fgithub.com\u002Fgoogle\u002Fadk-python\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fadk-debug\u002Freferences\u002Fcli-run.md",{"path":85,"resourceType":80,"resourceTypeDesc":81,"sizeBytes":86,"sourceHtmlUrl":87},"references\u002Fevent-flow.md",4681,"https:\u002F\u002Fgithub.com\u002Fgoogle\u002Fadk-python\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fadk-debug\u002Freferences\u002Fevent-flow.md",{"path":89,"resourceType":80,"resourceTypeDesc":81,"sizeBytes":90,"sourceHtmlUrl":91},"references\u002Ffailure-modes.md",5400,"https:\u002F\u002Fgithub.com\u002Fgoogle\u002Fadk-python\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fadk-debug\u002Freferences\u002Ffailure-modes.md",{"path":93,"resourceType":80,"resourceTypeDesc":81,"sizeBytes":94,"sourceHtmlUrl":95},"references\u002Flogs-and-traces.md",4352,"https:\u002F\u002Fgithub.com\u002Fgoogle\u002Fadk-python\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fadk-debug\u002Freferences\u002Flogs-and-traces.md",{"path":97,"resourceType":80,"resourceTypeDesc":81,"sizeBytes":98,"sourceHtmlUrl":99},"references\u002Fweb-api.md",3505,"https:\u002F\u002Fgithub.com\u002Fgoogle\u002Fadk-python\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fadk-debug\u002Freferences\u002Fweb-api.md",[101,130,156,182,198,206],{"skillNo":102,"name":103,"nameZh":6,"description":104,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":105,"categories":106,"keywords":111,"publishedAt":49},"01M698ZJNXJCE7T81AK0Y5R01D","bigquery-graph","Skill for Graph Query Language (GQL) or SQL\u002FPGQ queries against a property graph.\nIncludes path finding, multi-hop traversal, topological connection, shortest path,\nnode reachability, edge connectivity, and semantic graph queries.","https:\u002F\u002Fgithub.com\u002Fgoogle\u002Fadk-python\u002Fblob\u002Fmain\u002Fsrc\u002Fgoogle\u002Fadk\u002Fintegrations\u002Fbigquery\u002Fskills\u002Fbigquery-graph\u002FSKILL.md",[107,108],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},{"categoryNo":109,"categoryName":110,"taxonomyType":26,"taxonomyTypeDesc":27},"01JBSEEDCAT000000000000033","数据分析师",[112,113,114,9,115,30,116,117,118,119,120,121,122,123,124,125,126,127,128,129],"graph","bigquery","bigquery graph","for","sql","edge","graph query","hop","node","queries","language","query","connectivity","gql","semantic","multi","connection","path",{"skillNo":131,"name":132,"nameZh":6,"description":133,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":134,"categories":135,"keywords":137,"publishedAt":49},"01M698ZEVMRDDFX2MH3GHQGSNW","bigquery-ai-ml","Leverages BigQuery's built-in machine learning and GenAI capabilities for advanced data analytics. Use when you need to write SQL queries that perform time-series forecasting, predict values, detect outliers or anomalies, find key drivers, perform semantic search or vector search, classify text, calculate similarity, summarize content, translate language, evaluate models, filter by semantic conditions, measure the causal effect of an intervention, compute correlations between columns, detect change points or structural breaks, extract trend or seasonality components, or leverage generative AI capabilities in BigQuery. Do not use for general BigQuery dataset, table, or job management requests.","https:\u002F\u002Fgithub.com\u002Fgoogle\u002Fadk-python\u002Fblob\u002Fmain\u002Fsrc\u002Fgoogle\u002Fadk\u002Fintegrations\u002Fbigquery\u002Fskills\u002Fbigquery-ai-ml\u002FSKILL.md",[136],{"categoryNo":109,"categoryName":110,"taxonomyType":26,"taxonomyTypeDesc":27},[138,113,139,140,141,32,142,35,143,144,145,146,147,148,149,9,150,151,30,152,153,154,155],"ai ml","ml","bigquery ai ml","vector","generative","advanced","dataset","leverage","structural","time","effect","search","summarize","genai","generative ai","translate","components","capabilities",{"skillNo":157,"name":158,"nameZh":6,"description":159,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":160,"categories":161,"keywords":167,"publishedAt":49},"01M698ZCN6SZ9SJ8E9970M47FQ","code-review-skill","Reviews code with feedback tailored to the developer's profile in session state.","https:\u002F\u002Fgithub.com\u002Fgoogle\u002Fadk-python\u002Fblob\u002Fmain\u002Fcontributing\u002Fsamples\u002Fenvironment_and_skills\u002Fskills_inject_state\u002Fskills\u002Fcode-review-skill\u002FSKILL.md",[162,165,166],{"categoryNo":163,"categoryName":164,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000007","代码审查",{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},{"categoryNo":24,"categoryName":25,"taxonomyType":26,"taxonomyTypeDesc":27},[168,169,170,171,172,173,174,175,176,177,178,179,180,181],"review skill","code","code review skill","review","code review","session","state","developer","feedback","profile","with","session state","reviews","true",{"skillNo":183,"name":184,"nameZh":6,"description":185,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":186,"categories":187,"keywords":189,"publishedAt":49},"01M698ZAMH02NSGXMR533ETWYW","weather-skill","A skill that provides weather information based on reference data.","https:\u002F\u002Fgithub.com\u002Fgoogle\u002Fadk-python\u002Fblob\u002Fmain\u002Fcontributing\u002Fsamples\u002Fenvironment_and_skills\u002Fskills_agent\u002Fskills\u002Fweather-skill\u002FSKILL.md",[188],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},[190,191,192,193,194,195,196,197],"weather skill","weather","data","reference","speed","get","wind","information",{"skillNo":199,"name":184,"nameZh":6,"description":200,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":201,"categories":202,"keywords":204,"publishedAt":49},"01M698Z8QFDZ713E03589BHEG5","A skill that provides weather information based on reference data and scripts.","https:\u002F\u002Fgithub.com\u002Fgoogle\u002Fadk-python\u002Fblob\u002Fmain\u002Fcontributing\u002Fsamples\u002Fenvironment_and_skills\u002Fskills\u002Fskills\u002Fweather-skill\u002FSKILL.md",[203],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},[191,190,197,196,205,195,194,193,192],"scripts",{"skillNo":207,"name":184,"nameZh":6,"description":185,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":208,"categories":209,"keywords":211,"publishedAt":49},"01M698Z6P1NZFW6A6V342AWZPR","https:\u002F\u002Fgithub.com\u002Fgoogle\u002Fadk-python\u002Fblob\u002Fmain\u002Fcontributing\u002Fsamples\u002Fenvironment_and_skills\u002Flocal_environment_skill\u002Fskills\u002Fweather-skill\u002FSKILL.md",[210],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},[190,191,193,197,192]]