[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M758KAWPBX6WTACY6GSYYAYX":3,"skill-same:01M758KAWPBX6WTACY6GSYYAYX":64},{"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":41,"commitSha":42,"skillDirectory":43,"forksCount":44,"repositoryHtmlUrl":45,"repositoryHomepageUrl":46,"ownerHtmlUrl":47,"ownerBlogUrl":6,"ownerDescription":48,"updatedAt":49,"install":50,"resources":63},"01M758KAWPBX6WTACY6GSYYAYX","watchdog-alert-design",null,"Fix periodic health watchdogs that alert without flapping.","lukemcqueen\u002Fsteadfaste-fleet","lukemcqueen","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F405240?v=4","01M757MHPAGRXTJAT46SKZP98R","Luke McQueen","01M757MHPZ573KP8DT9GB8SBQ7",4,"MIT","https:\u002F\u002Fgithub.com\u002Flukemcqueen\u002Fsteadfaste-fleet\u002Fblob\u002Fmain\u002Fskills\u002Fdevops\u002Fwatchdog-alert-design\u002FSKILL.md",[],[],"2026-11-12 13:55:43",false,"---\nname: watchdog-alert-design\ndescription: Fix periodic health watchdogs that alert without flapping.\nversion: 1.0.0\ncategory: devops\nauthor: Steadfaste Fleet\nlicense: MIT\nmetadata:\n  hermes:\n    tags: [watchdog, health, cron, alert, debounce, reachability]\n    related_skills: [fleet-bus, cron-quality-gate, approval-gate-debugging, watchdog-flapping-diagnostics]\n---\n\n# Watchdog Alert Design (health \u002F bus \u002F reachability crons)\n\n## When to Use\n\n- Writing a new periodic health \u002F bus \u002F reachability watchdog cron.\n- Debugging a recurring warn→recover→warn flapping pair from an existing watchdog.\n- Reviewing an alert cron that never stays quiet.\n\nA periodic no_agent checker that warns on state changes is only worth deploying if\nit stays quiet when nothing is actually wrong. Every rule below is a failure mode\nseen in real fleet watchdogs.\n\n## Rules\n\n### 1. Silent-when-clean — no output = no notification\nHealthy state produces ZERO stdout. Collect the whole report first, then print\nonly if there is a signal. Never emit \"✅ all clean\".\n\n### 2. Debounce BOTH transitions, not just one\n- Alert only after **N consecutive** failed probes.\n- Declare recovery only after **M consecutive** successful probes.\n\nA path that alerts on its FIRST failure and recovers on its FIRST success turns\nany single transient probe blip (one slow timeout, one nginx hiccup, one dropped\nconnection) into an **alert+recovery flapping pair** — recurring spam on every tick\nas the blip re-occurs. Works defaults that hold up: **N=2** for the alert,\n**M=3** for the recovery.\n\n### 3. Trace flapping to the branch missing the debounce\nWhen a watchdog flaps warn→recover→warn, the broken branch is the one with no\nconsecutive-count gate. Its sibling role\u002Fstate branch almost always has one\n(a standby orchestrator path already requires 3 failures + elapsed time before it\nacts) — mirror that sibling's threshold onto the under-debounced branch instead of\ndesigning a new number.\n\n### 4. Real outages still report (once)\nDebounce must not hide a genuine outage. With N consecutive failures the alert\nstill fires once and then stays quiet while the condition holds; recovery needs M\nconsecutive successes so a one-tick recovery doesn't instantly flip back. A\ndebounce (count to threshold) is the right mechanism; a bare `last_state` toggle\nis the bug.\n\n## Testing watchdogs — make it host-independent\n\nWatchdogs read role (orchestrator vs worker) and health-URL config from the host\nat import, so running the drill on a host where the role short-circuits the logic\n(Moses host → `IS_MOSES` returns silent) or where no health URLs are configured\n(empty probe list → config-guard stands down) proves nothing. To test the state\nmachine on ANY host:\n\n- Force the role: override the module's `IS_MOSES` \u002F `IS_ORCHESTRATOR` to the\n  non-actor (worker) value after import.\n- Inject probe results: pass booleans for \"primary up \u002F backup up\" into a\n  `run_once(primary_up=..., backup_up=...)`-style seam instead of letting it hit\nthe network — this also makes the test hermetic and fast.\n- Redirect the state file to a temp dir so the drill never touches live state.\n\nWrite the regression test FIRST (it should FAIL against the undebounced code:\nassert a single blip emits zero output), then fix, then watch it go green.\n\n## Pitfalls\n\n- **A single-tick blip is real.** Treat any \"unreachable→reachable\" pair in the\n  same cron delivery as the signature of a missing-debounce bug, not two separate\ngenuine events — the recovery line can only fire after the alert, and both come\nfrom the probe result flipping twice across consecutive ticks.\n- **Don't copy the alert threshold onto recovery.** They are independent: recovery\n  should be at least as strict (more consecutive successes) than the alert\n  threshold, so a blip that just crossed the alert line doesn't instantly clear it.\n- **Empty health-URL list is \"unconfigured\", not \"down\".** A probe function that\n  returns False for an empty\u002FNone URL list will scream \"outage\" for an unset env\n  var. Fail closed the other way: stand down silently when nothing is configured.\n\n## Seam note\n\nThe clean shape is a `run_once(...)`-style pure function that returns the report\nlines and takes injectable probe booleans (production: HTTP adapter; test: injected\nvalues). One real adapter + one test adapter justifies the seam; the module's I\u002FO\n(state file, config swap, notify) stays behind it so tests hit only the function.\n","# Watchdog Alert Design (health \u002F bus \u002F reachability crons)\n\n## When to Use\n\n- Writing a new periodic health \u002F bus \u002F reachability watchdog cron.\n- Debugging a recurring warn→recover→warn flapping pair from an existing watchdog.\n- Reviewing an alert cron that never stays quiet.\n\nA periodic no_agent checker that warns on state changes is only worth deploying if\nit stays quiet when nothing is actually wrong. Every rule below is a failure mode\nseen in real fleet watchdogs.\n\n## Rules\n\n### 1. Silent-when-clean — no output = no notification\nHealthy state produces ZERO stdout. Collect the whole report first, then print\nonly if there is a signal. Never emit \"✅ all clean\".\n\n### 2. Debounce BOTH transitions, not just one\n- Alert only after **N consecutive** failed probes.\n- Declare recovery only after **M consecutive** successful probes.\n\nA path that alerts on its FIRST failure and recovers on its FIRST success turns\nany single transient probe blip (one slow timeout, one nginx hiccup, one dropped\nconnection) into an **alert+recovery flapping pair** — recurring spam on every tick\nas the blip re-occurs. Works defaults that hold up: **N=2** for the alert,\n**M=3** for the recovery.\n\n### 3. Trace flapping to the branch missing the debounce\nWhen a watchdog flaps warn→recover→warn, the broken branch is the one with no\nconsecutive-count gate. Its sibling role\u002Fstate branch almost always has one\n(a standby orchestrator path already requires 3 failures + elapsed time before it\nacts) — mirror that sibling's threshold onto the under-debounced branch instead of\ndesigning a new number.\n\n### 4. Real outages still report (once)\nDebounce must not hide a genuine outage. With N consecutive failures the alert\nstill fires once and then stays quiet while the condition holds; recovery needs M\nconsecutive successes so a one-tick recovery doesn't instantly flip back. A\ndebounce (count to threshold) is the right mechanism; a bare `last_state` toggle\nis the bug.\n\n## Testing watchdogs — make it host-independent\n\nWatchdogs read role (orchestrator vs worker) and health-URL config from the host\nat import, so running the drill on a host where the role short-circuits the logic\n(Moses host → `IS_MOSES` returns silent) or where no health URLs are configured\n(empty probe list → config-guard stands down) proves nothing. To test the state\nmachine on ANY host:\n\n- Force the role: override the module's `IS_MOSES` \u002F `IS_ORCHESTRATOR` to the\n  non-actor (worker) value after import.\n- Inject probe results: pass booleans for \"primary up \u002F backup up\" into a\n  `run_once(primary_up=..., backup_up=...)`-style seam instead of letting it hit\nthe network — this also makes the test hermetic and fast.\n- Redirect the state file to a temp dir so the drill never touches live state.\n\nWrite the regression test FIRST (it should FAIL against the undebounced code:\nassert a single blip emits zero output), then fix, then watch it go green.\n\n## Pitfalls\n\n- **A single-tick blip is real.** Treat any \"unreachable→reachable\" pair in the\n  same cron delivery as the signature of a missing-debounce bug, not two separate\ngenuine events — the recovery line can only fire after the alert, and both come\nfrom the probe result flipping twice across consecutive ticks.\n- **Don't copy the alert threshold onto recovery.** They are independent: recovery\n  should be at least as strict (more consecutive successes) than the alert\n  threshold, so a blip that just crossed the alert line doesn't instantly clear it.\n- **Empty health-URL list is \"unconfigured\", not \"down\".** A probe function that\n  returns False for an empty\u002FNone URL list will scream \"outage\" for an unset env\n  var. Fail closed the other way: stand down silently when nothing is configured.\n\n## Seam note\n\nThe clean shape is a `run_once(...)`-style pure function that returns the report\nlines and takes injectable probe booleans (production: HTTP adapter; test: injected\nvalues). One real adapter + one test adapter justifies the seam; the module's I\u002FO\n(state file, config swap, notify) stays behind it so tests hit only the function.\n",{"name":5,"author":24,"license":15,"version":25,"category":26,"metadata":27,"description":7},"Steadfaste Fleet","1.0.0","devops",{"hermes":28},{"tags":29,"related_skills":36},[30,31,32,33,34,35],"watchdog","health","cron","alert","debounce","reachability",[37,38,39,40],"fleet-bus","cron-quality-gate","approval-gate-debugging","watchdog-flapping-diagnostics","main","162463c5db5a1e06f4c54abddce6ee13c3d7612b","skills\u002Fdevops\u002Fwatchdog-alert-design",1,"https:\u002F\u002Fgithub.com\u002Flukemcqueen\u002Fsteadfaste-fleet","steadfaste.com","https:\u002F\u002Fgithub.com\u002Flukemcqueen","AI Architect | Building Hermes Cortex: governed multi-agent AI for autonomous software engineering | Enterprise Tech Consultant","2026-11-14 22:33:34",{"source":8,"skillName":5,"commands":51,"prompt":61,"zipAvailable":62},[52,55,58],{"tool":53,"command":54},"npx","npx skills add lukemcqueen\u002Fsteadfaste-fleet --skill watchdog-alert-design",{"tool":56,"command":57},"bunx","bunx skills add lukemcqueen\u002Fsteadfaste-fleet --skill watchdog-alert-design",{"tool":59,"command":60},"pnpm","pnpm dlx skills add lukemcqueen\u002Fsteadfaste-fleet --skill watchdog-alert-design","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Flukemcqueen\u002Fsteadfaste-fleet\u002Fblob\u002Fmain\u002Fskills\u002Fdevops\u002Fwatchdog-alert-design\u002FSKILL.md",true,[],[65,73,80,87,94,101],{"skillNo":66,"name":67,"nameZh":6,"description":68,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":69,"categories":70,"keywords":71,"publishedAt":72},"01M7DXY4G8NSVB59K8B6S2AQNP","macbook-linux-maintenance","Legacy MacBook Linux fixes: GPU wedge, reboot hang, thermal.","https:\u002F\u002Fgithub.com\u002Flukemcqueen\u002Fsteadfaste-fleet\u002Fblob\u002Fmain\u002Fskills\u002Fdevops\u002Fmacbook-linux-maintenance\u002FSKILL.md",[],[],"2026-11-15 14:19:17",{"skillNo":74,"name":75,"nameZh":6,"description":76,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":77,"categories":78,"keywords":79,"publishedAt":72},"01M7DXXYJCXFJ3TGB84ZQ8KFE6","governance-regression-triage","Triage a host's governance issues: what changed.","https:\u002F\u002Fgithub.com\u002Flukemcqueen\u002Fsteadfaste-fleet\u002Fblob\u002Fmain\u002Fskills\u002Fdevops\u002Fgovernance-regression-triage\u002FSKILL.md",[],[],{"skillNo":81,"name":82,"nameZh":6,"description":83,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":84,"categories":85,"keywords":86,"publishedAt":72},"01M7DXXSAP1G8FDRJAMXQGBBMK","enforcement-immutability","Use when enforcement files lose their immutable flag.","https:\u002F\u002Fgithub.com\u002Flukemcqueen\u002Fsteadfaste-fleet\u002Fblob\u002Fmain\u002Fskills\u002Fdevops\u002Fenforcement-immutability\u002FSKILL.md",[],[],{"skillNo":88,"name":89,"nameZh":6,"description":90,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":91,"categories":92,"keywords":93,"publishedAt":72},"01M78SSDQVRYX64Y16X32TC0NS","steadfaste-fleet-repo-deploy","Use when committing to the steadfaste-fleet repo.","https:\u002F\u002Fgithub.com\u002Flukemcqueen\u002Fsteadfaste-fleet\u002Fblob\u002Fmain\u002Fskills\u002Fdevops\u002Fsteadfaste-fleet-repo-deploy\u002FSKILL.md",[],[],{"skillNo":95,"name":96,"nameZh":6,"description":97,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":98,"categories":99,"keywords":100,"publishedAt":72},"01M78SS928F3R2RJAYVW4E30CE","steadfaste-fleet-maintenance","Maintain an installed Steadfaste Fleet instance — update both the upstream Hermes Agent and the fleet repo layer, merge diverged histories, sync skills safely, and troubleshoot common issues. Covers the dual-update workflow (hermes update + fleet-update.sh), encrypted brain data backup, mycortex import, and recovery techniques.","https:\u002F\u002Fgithub.com\u002Flukemcqueen\u002Fsteadfaste-fleet\u002Fblob\u002Fmain\u002Fskills\u002Fdevops\u002Fsteadfaste-fleet-maintenance\u002FSKILL.md",[],[],{"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":107,"publishedAt":72},"01M78SP1CHW0T0BWE5ZT6JBPX8","fleet-preflight","DEPRECATED alias — merged into survey-before-action (2026-08-20). Load survey-before-action instead; its \"Repo-Specific Pre-Flight\" section covers git search, Hermes boundary, and deployment verification.","https:\u002F\u002Fgithub.com\u002Flukemcqueen\u002Fsteadfaste-fleet\u002Fblob\u002Fmain\u002Fskills\u002Fdevops\u002Ffleet-preflight\u002FSKILL.md",[],[]]