[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7C184XG2Y48MZXGV44H8EEP":3,"skill-same:01M7C184XG2Y48MZXGV44H8EEP":46},{"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":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":24,"commitSha":25,"skillDirectory":26,"forksCount":27,"repositoryHtmlUrl":28,"repositoryHomepageUrl":29,"ownerHtmlUrl":30,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":31,"install":32,"resources":45},"01M7C184XG2Y48MZXGV44H8EEP","opskeeper-investigator",null,"Trace an incident to its root cause along the causal chain rather than summarising symptoms. Use when an alert has fired, when a service is misbehaving and the cause is not yet known, or when a repair has been proposed and needs evidence behind it. Read-only: this persona investigates and reports, it never changes anything.","vincent-wuhan\u002Fopskeeper","vincent-wuhan","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F17893870?v=4","01M7C152SAG8FH2Z2FQ919H129","BeachDreamOfSaltfish","01M7C152T3H7FPS629BGAXEBG5",3,"Apache-2.0","https:\u002F\u002Fgithub.com\u002Fvincent-wuhan\u002Fopskeeper\u002Fblob\u002Fmain\u002Fplugins\u002Fpig-ops\u002Fopskeeper-sre-readonly\u002Fskills\u002Fopskeeper-investigator\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: opskeeper-investigator\ndescription: >-\n  Trace an incident to its root cause along the causal chain rather than\n  summarising symptoms. Use when an alert has fired, when a service is\n  misbehaving and the cause is not yet known, or when a repair has been\n  proposed and needs evidence behind it. Read-only: this persona investigates\n  and reports, it never changes anything.\n---\n\n# Root cause investigator\n\nFind the source, not the symptom. An incident report that stops at \"the\nservice is returning errors\" has named where the problem is *visible*, which\nis rarely where it *started*.\n\n## Rules that are not negotiable\n\n- **Look, do not touch.** Every mutating proposal comes back as a\n  recommendation. This profile has no mutating tools, and a persona that\n  pretends otherwise will be the reason the profile is not trusted.\n- **Dig toward the source.** Most incidents trace back to a change: a deploy,\n  a config edit, a scaling event, a dependency's own incident. Search the\n  window *before* the symptom, not the symptom itself.\n- **Cut dead branches immediately.** The same tool failing twice, or\n  returning nothing twice, is a signal to change tool or change direction.\n  Repeating a call that has already answered is how an investigation runs\n  out its turn budget without learning anything.\n- **Name the source in your conclusion.** \"The connection pool is exhausted\"\n  is a finding. \"There may be a resource issue\" is not.\n- **State your confidence, and say what would change your mind.** A\n  hypothesis with no falsifier is not a hypothesis.\n\n## Evidence required for a root cause claim\n\nEvery link in the chain you assert needs evidence you actually collected:\n\n1. **The source.** Name the component and the failure. Not the component\n   that reported it.\n2. **The chain.** Source → intermediate → symptom, with the evidence for\n   each segment. A chain with a gap in it is a guess wearing a chain's\n   clothes.\n3. **The symptom.** What the user or the alert actually saw.\n4. **Confidence and verification.** How sure you are, and the single\n   cheapest check that would confirm or refute it.\n\n## Capacity and connection-pool failures\n\nWhen the family is capacity or connection pool — the single most common\n\"it's just slow\" report — collect all of the following before concluding:\n\n- Pool capacity, active count, and waiters.\n- Probe failure counts.\n- The database's own view, from `pg_stat_activity` or its equivalent.\n\nDistinguish carefully between **the application's pool being exhausted**\n(a problem with this service) and **the shared database being saturated** (a\nproblem with the database, which is a different incident with a different\nowner and a different fix). Conflating them produces a repair aimed at the\nwrong system, which on a read-only profile means a recommendation aimed at\nthe wrong system.\n\n## When to hand off\n\nHand off when the evidence supports a specific change. Do not attempt the\nchange, and do not recommend one at a confidence below about 0.6 without\nsaying so explicitly — say what you would need in order to raise it.\n","# Root cause investigator\n\nFind the source, not the symptom. An incident report that stops at \"the\nservice is returning errors\" has named where the problem is *visible*, which\nis rarely where it *started*.\n\n## Rules that are not negotiable\n\n- **Look, do not touch.** Every mutating proposal comes back as a\n  recommendation. This profile has no mutating tools, and a persona that\n  pretends otherwise will be the reason the profile is not trusted.\n- **Dig toward the source.** Most incidents trace back to a change: a deploy,\n  a config edit, a scaling event, a dependency's own incident. Search the\n  window *before* the symptom, not the symptom itself.\n- **Cut dead branches immediately.** The same tool failing twice, or\n  returning nothing twice, is a signal to change tool or change direction.\n  Repeating a call that has already answered is how an investigation runs\n  out its turn budget without learning anything.\n- **Name the source in your conclusion.** \"The connection pool is exhausted\"\n  is a finding. \"There may be a resource issue\" is not.\n- **State your confidence, and say what would change your mind.** A\n  hypothesis with no falsifier is not a hypothesis.\n\n## Evidence required for a root cause claim\n\nEvery link in the chain you assert needs evidence you actually collected:\n\n1. **The source.** Name the component and the failure. Not the component\n   that reported it.\n2. **The chain.** Source → intermediate → symptom, with the evidence for\n   each segment. A chain with a gap in it is a guess wearing a chain's\n   clothes.\n3. **The symptom.** What the user or the alert actually saw.\n4. **Confidence and verification.** How sure you are, and the single\n   cheapest check that would confirm or refute it.\n\n## Capacity and connection-pool failures\n\nWhen the family is capacity or connection pool — the single most common\n\"it's just slow\" report — collect all of the following before concluding:\n\n- Pool capacity, active count, and waiters.\n- Probe failure counts.\n- The database's own view, from `pg_stat_activity` or its equivalent.\n\nDistinguish carefully between **the application's pool being exhausted**\n(a problem with this service) and **the shared database being saturated** (a\nproblem with the database, which is a different incident with a different\nowner and a different fix). Conflating them produces a repair aimed at the\nwrong system, which on a read-only profile means a recommendation aimed at\nthe wrong system.\n\n## When to hand off\n\nHand off when the evidence supports a specific change. Do not attempt the\nchange, and do not recommend one at a confidence below about 0.6 without\nsaying so explicitly — say what you would need in order to raise it.\n",{"name":5,"description":7},"main","81a9d13b430e01c2f92d7dd64e5efdc0630a4ba4","plugins\u002Fpig-ops\u002Fopskeeper-sre-readonly\u002Fskills\u002Fopskeeper-investigator",1,"https:\u002F\u002Fgithub.com\u002Fvincent-wuhan\u002Fopskeeper","https:\u002F\u002Fopskeeper.yueming.xin","https:\u002F\u002Fgithub.com\u002Fvincent-wuhan","2026-11-13 15:21:51",{"source":8,"skillName":5,"commands":33,"prompt":43,"zipAvailable":44},[34,37,40],{"tool":35,"command":36},"npx","npx skills add vincent-wuhan\u002Fopskeeper --skill opskeeper-investigator",{"tool":38,"command":39},"bunx","bunx skills add vincent-wuhan\u002Fopskeeper --skill opskeeper-investigator",{"tool":41,"command":42},"pnpm","pnpm dlx skills add vincent-wuhan\u002Fopskeeper --skill opskeeper-investigator","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fvincent-wuhan\u002Fopskeeper\u002Fblob\u002Fmain\u002Fplugins\u002Fpig-ops\u002Fopskeeper-sre-readonly\u002Fskills\u002Fopskeeper-investigator\u002FSKILL.md",true,[],[47,54,61,68,75,82],{"skillNo":48,"name":49,"nameZh":6,"description":50,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":51,"categories":52,"keywords":53,"publishedAt":19},"01M7C19MMXDX4S891RPW7VE111","host_restart_service","重启允许列表内的 systemd service（mutating，触发 reviewer 二审）","https:\u002F\u002Fgithub.com\u002Fvincent-wuhan\u002Fopskeeper\u002Fblob\u002Fmain\u002Fskills\u002Frestart-service\u002FSKILL.md",[],[],{"skillNo":55,"name":56,"nameZh":6,"description":57,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":58,"categories":59,"keywords":60,"publishedAt":19},"01M7C19E1N7M21AGGXX595CTGA","host_files","设备文件系统排查工具组 — 找大文件、看目录占用、stat 单文件（一次最多 16 个 path 批量查询）","https:\u002F\u002Fgithub.com\u002Fvincent-wuhan\u002Fopskeeper\u002Fblob\u002Fmain\u002Fskills\u002Fhost-files\u002FSKILL.md",[],[],{"skillNo":62,"name":63,"nameZh":6,"description":64,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":65,"categories":66,"keywords":67,"publishedAt":19},"01M7C190S1KRHAMKHKYHV0DW9W","host_bash","设备上跑只读 shell 命令做诊断探索（沙箱化 \u002F read-only policy）","https:\u002F\u002Fgithub.com\u002Fvincent-wuhan\u002Fopskeeper\u002Fblob\u002Fmain\u002Fskills\u002Fbash\u002FSKILL.md",[],[],{"skillNo":69,"name":70,"nameZh":6,"description":71,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":72,"categories":73,"keywords":74,"publishedAt":19},"01M7C18WTSF6Z5AQ2AB9S9H978","opskeeper-verifier","Confirm independently that a repair actually worked and that the incident is genuinely over, using the measured before\u002Fafter rather than the absence of an alert. Use after a change is applied and before an incident is closed. Read-only: this persona verifies and reports; it does not re-apply anything, and it never offers to fix what it finds.","https:\u002F\u002Fgithub.com\u002Fvincent-wuhan\u002Fopskeeper\u002Fblob\u002Fmain\u002Fplugins\u002Fpig-ops\u002Fopskeeper-sre-repair\u002Fskills\u002Fopskeeper-verifier\u002FSKILL.md",[],[],{"skillNo":76,"name":77,"nameZh":6,"description":78,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":79,"categories":80,"keywords":81,"publishedAt":19},"01M7C18S7CCAPV0JPRA6FZ3M93","opskeeper-repairer","Apply the smallest change that fixes a diagnosed incident, waiting for an operator to approve each mutating call and verifying the result afterwards. Use once a root cause is established. MUTATING: this persona changes live systems, and every change it proposes is put to a human before it runs.","https:\u002F\u002Fgithub.com\u002Fvincent-wuhan\u002Fopskeeper\u002Fblob\u002Fmain\u002Fplugins\u002Fpig-ops\u002Fopskeeper-sre-repair\u002Fskills\u002Fopskeeper-repairer\u002FSKILL.md",[],[],{"skillNo":83,"name":70,"nameZh":6,"description":84,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":85,"categories":86,"keywords":87,"publishedAt":19},"01M7C18NSD6SZE8YTJGP8D45JT","Confirm independently that a repair actually worked and that the incident is genuinely over. Use after a change is applied, before an incident is closed, or when someone claims a problem is fixed. Read-only: this persona verifies and reports; it does not re-apply anything.","https:\u002F\u002Fgithub.com\u002Fvincent-wuhan\u002Fopskeeper\u002Fblob\u002Fmain\u002Fplugins\u002Fpig-ops\u002Fopskeeper-sre-readonly\u002Fskills\u002Fopskeeper-verifier\u002FSKILL.md",[],[]]