[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M72SWTS25M60M51W9HYZDFWM":3,"skill-same:01M72SWTS25M60M51W9HYZDFWM":62},{"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":35,"commitSha":36,"skillDirectory":37,"forksCount":14,"repositoryHtmlUrl":38,"repositoryHomepageUrl":38,"ownerHtmlUrl":39,"ownerBlogUrl":6,"ownerDescription":40,"updatedAt":41,"install":42,"resources":55},"01M72SWTS25M60M51W9HYZDFWM","cron-fleet-maintenance",null,"Use when auditing or repairing a fleet of cron jobs.","iberi22\u002Fskills","iberi22","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F10615454?v=4","01M72SQ2NXAA3GW0FSH053A5ES","Brahyan Belalcazar","01M72SQ2PYCH0SSA1K61NGEHDV",0,"MIT","https:\u002F\u002Fgithub.com\u002Fiberi22\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fdevops\u002Fcron-fleet-maintenance\u002FSKILL.md",[],[],"2026-11-10 22:39:42",false,"---\nname: cron-fleet-maintenance\ndescription: \"Use when auditing or repairing a fleet of cron jobs.\"\nversion: 1.0.0\nauthor: Hermes Agent\nlicense: MIT\ntags: [cron, automation, fleet, audit, triage, no-agent, harness, cost]\n---\n\n# Cron Fleet Maintenance\n\nA fleet of scheduled jobs is a system, not a list. This is the class-level\nprocedure for auditing it, deciding what to keep, and repairing what is broken.\nApplies to any scheduler with the same primitives: a job registry file, per-job\nschedule\u002Fprompt\u002Fscript, and a `no_agent` (script-only) mode that costs no tokens.\n\nTriggers: audit my cron jobs; which jobs cost tokens; which jobs are obsolete;\nunify duplicate jobs; a job reports ok but fails; a scheduled job broke; job\nfails with 402 \u002F insufficient funds \u002F command not found.\n\n## Step 1 — Inventory before touching anything\n\nBack up first, then classify. Nothing else matters until you can answer these\nthree per job:\n\n1. **Does it cost tokens?** `no_agent: true` = script only, zero inference, no\n   matter how often it fires. Otherwise it costs once per tick — unless it is\n   monitor-gated (a script whose hash gates the wake-up), in which case it only\n   costs when that hash changes. A monitor-gated job idle for days is spending\n   nothing; say so rather than counting executions as cost.\n2. **Is it broken, or starved?** HTTP 402 \u002F \"insufficient funds\" means the\n   *provider* is out of credit, not that the job is defective. Starved jobs are\n   still worth keeping — pause, never delete. Genuinely obsolete means: script\n   deleted, workdir gone, or never executed. Verify both on disk before deciding.\n3. **Is it covered elsewhere?** Two jobs for one repo is duplication; two jobs\n   for two repos is not. Before pausing a \"redundant\" job, confirm the\n   replacement covers the *same* target. Redundancy you assumed is a coverage\n   gap you just created.\n\nRead the schedule field as a **structured dict**, not a cron string, and compute\nfrequency: `{\"kind\":\"interval\",\"minutes\":20}` is 72\u002Fday, `\"*\u002F30 * * * *\"` is\n48\u002Fday. Parsing it as a string yields nonsense (a 20-minute job does not fire\nonce a day).\n\n## Step 2 — Repair wrappers before trusting any status\n\nA job reporting `ok` can be lying. These three defects each made a job report\nsuccess while failing, and all are invisible from the registry:\n\n- **A pipe to a truncating command returns the wrong exit code.** `cmd | tail -n 8`\n  exits with *tail's* status (0), so any real failure reads as success. Use\n  `set -o pipefail` and exit with `${PIPESTATUS[0]}`. The identical shape:\n  `git add a b c 2>\u002Fdev\u002Fnull` aborts (rc 128) and stages **nothing** when one path\n  is missing, leaving any guard reading `git diff --cached` silently scanning an\n  empty set. Never suppress stderr on a command whose failure changes what you\n  inspect.\n- **A minimal env breaks absolute-looking assumptions.** Cron\u002Fsystemd env is not\n  your shell's. On NixOS `\u002Fusr\u002Fbin` does not exist and coreutils plus node live\n  under the system store path, so an exported PATH without it yields\n  \"command not found\" (exit 127) on the first utility the script uses.\n- **A script that reads a label\u002Fmarker that does not exist.** `gh issue create\n  --label X` does *not* create label X: gh fails, the issue goes out unlabelled,\n  and the script dies after having already written. Ensure the prerequisite is\n  present, idempotently, before the tool that requires it.\n\nTest every wrapper under an explicit minimal env, not your shell:\n\n```bash\nenv -i HOME=\"$HOME\" PATH=\"\u002Frun\u002Fcurrent-system\u002Fsw\u002Fbin:$HOME\u002F.local\u002Fbin:\u002Fusr\u002Fbin:\u002Fbin\" \\\n  bash wrapper.sh; echo \"EXIT=$?\"\n```\n\n## Step 3 — Verify by the real channel\n\nShell output is not the job's environment; the dispatcher's own session is.\nTrigger the job and let it run where it actually runs, then read back\n`last_status`\u002F`last_error`. A job that only ever passed in your shell is unproven.\n\n## Always-on rules\n\n- **Back up in two artifacts before any destructive step**: the full prior state\n  (`registry.json.bak`) and a manifest of `id + reason` per removed entry.\n  \"Reversible\" means another session can restore without reconstruction.\n- **Prefer pause over delete** for anything that might be needed again (starved\n  providers, seasonal work). Delete only after confirming on disk that the\n  script and workdir are gone.\n- **Clean residual state on conversion.** When a job goes `no_agent`, clear its\n  `model`\u002F`provider`\u002F`skills`\u002F`prompt`. They are inert today but are a trap: the\n  day someone removes `no_agent`, the job silently reverts to a provider that may\n  have no credit. State should reflect what the job does.\n- **Report cost in executions\u002Fday, split by whether inference is involved** —\n  that split is the whole reason the audit is worth running.\n- **State what you did not verify.** Jobs that have never fired (future-dated),\n  repos left without coverage, or jobs skipped for lack of credit belong in an\n  explicit \"pending\" list, not quietly in a green table.\n\n## Verifying when no test command exists\n\nMost cron wrappers have no repo, no CI, no canonical gate. \"No gate exists\" means\nyou build one — and a harness you delete at the end of the turn guarantees the\nsame verification gets rebuilt next session. Full procedure, plus the\n`git add` \u002F `git rm --cached` staging traps that bite staging guards, in\n`references\u002Fadhoc-harness-and-staging-traps.md`.\n\n## Support files\n\n- `references\u002Fadhoc-harness-and-staging-traps.md` — building an ad-hoc harness\n  with checks that have teeth, and the git-index behaviours that make\n  pre-commit\u002Fstaging guards silently pass\n","# Cron Fleet Maintenance\n\nA fleet of scheduled jobs is a system, not a list. This is the class-level\nprocedure for auditing it, deciding what to keep, and repairing what is broken.\nApplies to any scheduler with the same primitives: a job registry file, per-job\nschedule\u002Fprompt\u002Fscript, and a `no_agent` (script-only) mode that costs no tokens.\n\nTriggers: audit my cron jobs; which jobs cost tokens; which jobs are obsolete;\nunify duplicate jobs; a job reports ok but fails; a scheduled job broke; job\nfails with 402 \u002F insufficient funds \u002F command not found.\n\n## Step 1 — Inventory before touching anything\n\nBack up first, then classify. Nothing else matters until you can answer these\nthree per job:\n\n1. **Does it cost tokens?** `no_agent: true` = script only, zero inference, no\n   matter how often it fires. Otherwise it costs once per tick — unless it is\n   monitor-gated (a script whose hash gates the wake-up), in which case it only\n   costs when that hash changes. A monitor-gated job idle for days is spending\n   nothing; say so rather than counting executions as cost.\n2. **Is it broken, or starved?** HTTP 402 \u002F \"insufficient funds\" means the\n   *provider* is out of credit, not that the job is defective. Starved jobs are\n   still worth keeping — pause, never delete. Genuinely obsolete means: script\n   deleted, workdir gone, or never executed. Verify both on disk before deciding.\n3. **Is it covered elsewhere?** Two jobs for one repo is duplication; two jobs\n   for two repos is not. Before pausing a \"redundant\" job, confirm the\n   replacement covers the *same* target. Redundancy you assumed is a coverage\n   gap you just created.\n\nRead the schedule field as a **structured dict**, not a cron string, and compute\nfrequency: `{\"kind\":\"interval\",\"minutes\":20}` is 72\u002Fday, `\"*\u002F30 * * * *\"` is\n48\u002Fday. Parsing it as a string yields nonsense (a 20-minute job does not fire\nonce a day).\n\n## Step 2 — Repair wrappers before trusting any status\n\nA job reporting `ok` can be lying. These three defects each made a job report\nsuccess while failing, and all are invisible from the registry:\n\n- **A pipe to a truncating command returns the wrong exit code.** `cmd | tail -n 8`\n  exits with *tail's* status (0), so any real failure reads as success. Use\n  `set -o pipefail` and exit with `${PIPESTATUS[0]}`. The identical shape:\n  `git add a b c 2>\u002Fdev\u002Fnull` aborts (rc 128) and stages **nothing** when one path\n  is missing, leaving any guard reading `git diff --cached` silently scanning an\n  empty set. Never suppress stderr on a command whose failure changes what you\n  inspect.\n- **A minimal env breaks absolute-looking assumptions.** Cron\u002Fsystemd env is not\n  your shell's. On NixOS `\u002Fusr\u002Fbin` does not exist and coreutils plus node live\n  under the system store path, so an exported PATH without it yields\n  \"command not found\" (exit 127) on the first utility the script uses.\n- **A script that reads a label\u002Fmarker that does not exist.** `gh issue create\n  --label X` does *not* create label X: gh fails, the issue goes out unlabelled,\n  and the script dies after having already written. Ensure the prerequisite is\n  present, idempotently, before the tool that requires it.\n\nTest every wrapper under an explicit minimal env, not your shell:\n\n```bash\nenv -i HOME=\"$HOME\" PATH=\"\u002Frun\u002Fcurrent-system\u002Fsw\u002Fbin:$HOME\u002F.local\u002Fbin:\u002Fusr\u002Fbin:\u002Fbin\" \\\n  bash wrapper.sh; echo \"EXIT=$?\"\n```\n\n## Step 3 — Verify by the real channel\n\nShell output is not the job's environment; the dispatcher's own session is.\nTrigger the job and let it run where it actually runs, then read back\n`last_status`\u002F`last_error`. A job that only ever passed in your shell is unproven.\n\n## Always-on rules\n\n- **Back up in two artifacts before any destructive step**: the full prior state\n  (`registry.json.bak`) and a manifest of `id + reason` per removed entry.\n  \"Reversible\" means another session can restore without reconstruction.\n- **Prefer pause over delete** for anything that might be needed again (starved\n  providers, seasonal work). Delete only after confirming on disk that the\n  script and workdir are gone.\n- **Clean residual state on conversion.** When a job goes `no_agent`, clear its\n  `model`\u002F`provider`\u002F`skills`\u002F`prompt`. They are inert today but are a trap: the\n  day someone removes `no_agent`, the job silently reverts to a provider that may\n  have no credit. State should reflect what the job does.\n- **Report cost in executions\u002Fday, split by whether inference is involved** —\n  that split is the whole reason the audit is worth running.\n- **State what you did not verify.** Jobs that have never fired (future-dated),\n  repos left without coverage, or jobs skipped for lack of credit belong in an\n  explicit \"pending\" list, not quietly in a green table.\n\n## Verifying when no test command exists\n\nMost cron wrappers have no repo, no CI, no canonical gate. \"No gate exists\" means\nyou build one — and a harness you delete at the end of the turn guarantees the\nsame verification gets rebuilt next session. Full procedure, plus the\n`git add` \u002F `git rm --cached` staging traps that bite staging guards, in\n`references\u002Fadhoc-harness-and-staging-traps.md`.\n\n## Support files\n\n- `references\u002Fadhoc-harness-and-staging-traps.md` — building an ad-hoc harness\n  with checks that have teeth, and the git-index behaviours that make\n  pre-commit\u002Fstaging guards silently pass\n",{"name":5,"tags":24,"author":33,"license":15,"version":34,"description":7},[25,26,27,28,29,30,31,32],"cron","automation","fleet","audit","triage","no-agent","harness","cost","Hermes Agent","1.0.0","main","688810f2886d212632acf81e532b24e483dd8b98","skills\u002Fdevops\u002Fcron-fleet-maintenance","https:\u002F\u002Fgithub.com\u002Fiberi22\u002Fskills","https:\u002F\u002Fgithub.com\u002Fiberi22","Senior WEB Full-Stack","2026-10-12 02:46:21",{"source":8,"skillName":5,"commands":43,"prompt":53,"zipAvailable":54},[44,47,50],{"tool":45,"command":46},"npx","npx skills add iberi22\u002Fskills --skill cron-fleet-maintenance",{"tool":48,"command":49},"bunx","bunx skills add iberi22\u002Fskills --skill cron-fleet-maintenance",{"tool":51,"command":52},"pnpm","pnpm dlx skills add iberi22\u002Fskills --skill cron-fleet-maintenance","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fiberi22\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fdevops\u002Fcron-fleet-maintenance\u002FSKILL.md",true,[56],{"path":57,"resourceType":58,"resourceTypeDesc":59,"sizeBytes":60,"sourceHtmlUrl":61},"references\u002Fadhoc-harness-and-staging-traps.md","REFERENCE","参考文档",5997,"https:\u002F\u002Fgithub.com\u002Fiberi22\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fdevops\u002Fcron-fleet-maintenance\u002Freferences\u002Fadhoc-harness-and-staging-traps.md",[63,69,76,82,88,94],{"skillNo":64,"name":65,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":66,"categories":67,"keywords":68,"publishedAt":19},"01M72TWQEQD46ZC9TRBAABF78P","web-design-guidelines","https:\u002F\u002Fgithub.com\u002Fiberi22\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fweb\u002Fweb-design-guidelines\u002FSKILL.md",[],[],{"skillNo":70,"name":71,"nameZh":6,"description":72,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":73,"categories":74,"keywords":75,"publishedAt":19},"01M72TWMQFTYD08DGXVAXRPPEB","blocked-page-recovery","Use when a fetch fails: 403\u002F429, paywall, WAF, bot wall. Recovers the page via archives and provenance-checked fallbacks.","https:\u002F\u002Fgithub.com\u002Fiberi22\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Fweb\u002Fblocked-page-recovery\u002FSKILL.md",[],[],{"skillNo":77,"name":78,"nameZh":6,"description":6,"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},"01M72TVQMS3HCTX18C8T97WTQQ","studio-coach","https:\u002F\u002Fgithub.com\u002Fiberi22\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Ftools\u002Fstudio-coach\u002FSKILL.md",[],[],{"skillNo":83,"name":84,"nameZh":6,"description":6,"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},"01M72TVNA2PR9RZW0ZAXYP1G6J","src-generator","https:\u002F\u002Fgithub.com\u002Fiberi22\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Ftools\u002Fsrc-generator\u002FSKILL.md",[],[],{"skillNo":89,"name":90,"nameZh":6,"description":6,"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":19},"01M72TVEJHHB6SAT3R3TY9MJKS","sqlite-pro","https:\u002F\u002Fgithub.com\u002Fiberi22\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Ftools\u002Fsqlite-pro\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":19},"01M72TVC2EVBHQHCX96VTTYKB8","rigor-evaluator","Use when you need a quick heuristic rigor score (10 criteria, 0-2 each) for the features in a .gitcore\u002Ffeatures.json file, to decide which features are ready to ship.","https:\u002F\u002Fgithub.com\u002Fiberi22\u002Fskills\u002Fblob\u002Fmain\u002Fskills\u002Ftools\u002Frigor-evaluator\u002FSKILL.md",[],[]]