[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7EMD4M0HT10W8AB5ENV2Z47":3,"skill-same:01M7EMD4M0HT10W8AB5ENV2Z47":59},{"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":14,"repositoryHtmlUrl":27,"repositoryHomepageUrl":6,"ownerHtmlUrl":28,"ownerBlogUrl":29,"ownerDescription":6,"updatedAt":30,"install":31,"resources":44},"01M7EMD4M0HT10W8AB5ENV2Z47","host-hygiene",null,"Plan and apply reviewed, exact-entry disk reclamation on Linux SSH hosts while protecting live and paused lanes, Docker artifacts, Nix roots, and shared temporary files.","tedks\u002Fagent-skills","tedks","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F1703535?v=4","01M7EMCJYBNAVD6EN62JAPCK0X","Ted Smith","01M7EMCJYZVVBFQTGZX8W2BXVT",0,"AGPL-3.0","https:\u002F\u002Fgithub.com\u002Ftedks\u002Fagent-skills\u002Fblob\u002Fmaster\u002Fskills\u002Fhost-hygiene\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: host-hygiene\ndescription: Plan and apply reviewed, exact-entry disk reclamation on Linux SSH hosts while protecting live and paused lanes, Docker artifacts, Nix roots, and shared temporary files.\n---\n\nFollow orgs `doctrine\u002FDOCTRINE.md`, “Host resources and hygiene”. The owning\nlane inventories and reviews; a **fresh Utility agent** receives exactly that\nreviewed list and applies it. Authorization for hygiene does not authorize\nkilling peers, deleting worktrees or peer caches, or changing shared tags.\n\nUse `scripts\u002Fhygiene.sh plan HOST PLAN.json --policy POLICY.json`. Linux,\nPython 3, SSH BatchMode, Docker and Nix tools are expected; nothing uses sudo.\nThe per-host policy must specify an absolute `projects_root` on the remote host.\nDo not substitute the local user's home-expanded root: another host or SSH user\nmay keep projects elsewhere. The [shared configuration](..\u002F_shared\u002Fconfiguration.md)\nCLI can supply a candidate only after you confirm that path on the target host.\nThe script runs over SSH without installing or creating remote helper files.\nUse a private local state directory, e.g. `~\u002F.local\u002Fstate\u002Fhygiene\u002FYYYYMMDD\u002FHOST\u002F`.\nPlan files are created exclusively with mode 0600. The printed SHA256 binds the\nwhole reviewed list; neither edit the list nor substitute a new inventory after\nreview. Before deletion, inspect each full ID\u002Fname\u002Fpath, size in bytes, ownership,\nreason, eligibility, and every diagnostic. Missing sizes are never deletable.\nBuildx versions may render human sizes: `size_display` preserves them,\n`size_exact=false` marks the `bytes` value as an estimate rather than an exact\nbyte measurement. Unsupported size formats disable cache eligibility. `\u002Ftmp`\nuses allocated inode blocks for nondirectories. Directories use GNU `du` in\nbatches of at most 256 exact paths, with NUL input\u002Foutput and a shared 60s\nmeasurement budget. Symlinks are not followed; path aliases, mounted roots or directories\ncontaining mounts are retained without traversal. Directory sizes count every\nhardlink to avoid cross-operand deduplication and are marked `size_exact=false`:\nthey can exceed an independent per-root deduplicated measurement and do not\npredict freed bytes. Missing\u002Funsupported output, stderr, nonzero exit, timeout,\nor changed directory\u002Fmount identity discards every size in that affected batch;\nall entries remain in the plan with unknown sizes and diagnostics. Failed batches\ninclude output byte counts and at most 512-byte excerpts per stream; subprocess\ncapture itself is not a streaming memory bound. CLI snapshots\ncannot exclude a concurrent mount or inode change atomically.\nMeasure `df -h` before and after separately and retain the plan and apply output.\n\nA small owner-maintained policy looks like this (replace full IDs):\n\n```json\n{\n  \"projects_root\": \"\u002Fabsolute\u002Fpath\u002Fto\u002Fprojects\",\n  \"records\": [\"\u002Fabsolute\u002Fpath\u002Fto\u002Fcurrent-owner-and-paused-lane-record.md\"],\n  \"protected\": [\"peer-service\"],\n  \"tmp_owner\": \"peer-lane\",\n  \"builder\": \"default\",\n  \"claims\": [\n    {\"kind\": \"container\", \"id\": \"FULL_64_HEX_ID\", \"owner\": \"my-lane\", \"idle\": true},\n    {\"kind\": \"image\", \"id\": \"sha256:FULL_64_HEX_ID\", \"owner\": \"my-lane\", \"idle\": true,\n     \"released_tags\": [\"my-owned-repo:old-tag\"]},\n    {\"kind\": \"volume\", \"id\": \"my-owned-compose-volume\", \"owner\": \"my-lane\", \"idle\": true},\n    {\"kind\": \"cache\", \"id\": \"FULL_BUILDKIT_CACHE_ID\", \"owner\": \"my-lane\", \"idle\": true}\n  ]\n}\n```\n\n`records` is the explicit, nonempty list of required local ownership\u002Freference\ndocuments. Include current ownership and pause records for every project on\nthe target host; confirm their paths with the owning lanes.\nThe owning lane confirms completeness; missing\u002Funreadable records disable all\neligibility. Incomplete remote project-root\u002Fcompose scans disable Docker eligibility\nwhile independent Nix dead-graph\u002Froot checks remain usable.\nReference scans have a 60s budget, reject symlink\u002Fnonregular compose files,\nand fail closed on symlink directories instead of silently missing external\nworktree references. Review external references explicitly and declare complete\ncoverage in the policy; do not blindly follow arbitrary links or loops.\nAn incomplete scan retains Docker\u002Fcache entries. Compose files below\n`projects_root`, including paused worktrees, conservatively protect references,\nexcept intentionally excluded `.git`, `.bare`, `node_modules` and `.venv` trees.\nThe owner must provide relevant compose\u002Flane references from excluded trees or\nother unscanned coverage in `records`; exclusions do not establish absence of\nreferences. Add other project roots' compose\u002Flane information as local\nrecords; the owning lane must confirm the declared root covers its host's worktrees.\n`protected` adds exact names, full IDs, tags or paths from peer leases and records.\nPaused is not abandoned: their images, volumes, unrooted store paths and caches\nmust be named in records or `protected`; do not grant their ownership to yourself.\n`claims` are affirmative owner-confirmed idle assertions, not inferred from age.\nConflicting claims for one resource, including a peer's non-idle claim, retain it;\none lane's idle claim cannot borrow another lane's tag-release authorization.\nFor images, `released_tags` is an explicit release of every current tag by its\nowner; shared tags may not be released by this lane.\n\nGive the fresh Utility agent the reviewed plan, digest, policy and owning lane:\n\n```sh\nscripts\u002Fhygiene.sh apply HOST PLAN.json --policy POLICY.json \\\n  --owner my-lane --reviewed-sha256 REVIEWED_DIGEST > APPLY.json\n```\n\nApply reloads current policy\u002Frecords and inventories the same host\u002Fuser, then\nmatches only saved eligible entries against fresh IDs, inspection fingerprints,\nages, references and ownership. It checks the exact Docker\u002Fbackend identity again\nbefore each deletion and rechecks the Nix dead graph before one exact-list daemon\ndelete batch (referrers and references can be deleted together).\nDocker removes exited containers older than 24h without force\u002Fvolume removal;\nuntagged unused images; tagged unused images older than 14d with explicit tag\nrelease; and unreferenced compose volumes whose size and mount identity can be\nread. No broad prune runs. Nix inventories roots and the dead graph with sizes;\nthe exact reviewed dead-path subset is passed as one batch to `nix-store --delete`, which\nrechecks liveness through the daemon. No roots\u002Fprofiles\u002Fgenerations are removed,\nand newly eligible paths are never added to the reviewed list. Store mtimes are\nnormalized and do not prove installation age; generation-age cleanup is deferred\nunder follow-up `host-hygiene-profile-age`. Keep paused\nunrooted closures protected explicitly; the daemon protects rooted closures.\n\nConcrete retained classes:\n\n- Build cache unused >72h is inventoried as NDJSON on the explicit `builder`.\n  Only immutable, private, reclaimable, owner-confirmed records qualify. Apply\n  pins the builder\u002Fbackend identity and uses ANDed `id=FULL_ID`, `until=72h`,\n  `private=true`, `inuse=false`, and `immutable=true` filters, then verifies the\n  selected ID is absent. Mutable\u002Fshared\u002Fpaused\u002Funowned entries are retained;\n  unsupported JSON or filter options fail closed without a broad-prune fallback.\n  See [Docker Buildx prune](https:\u002F\u002Fdocs.docker.com\u002Freference\u002Fcli\u002Fdocker\u002Fbuildx\u002Fprune\u002F)\n  and [disk usage](https:\u002F\u002Fdocs.docker.com\u002Freference\u002Fcli\u002Fdocker\u002Fbuildx\u002Fdu\u002F).\n- `\u002Ftmp` entries older than 2d are enumerated with exact paths and sizes, but no\n  `\u002Ftmp` deletion is implemented: unprivileged access cannot prove complete\n  process, mount and socket references across peers. Diagnostics name exact\n  `sudo lsof +D PATH` and `sudo findmnt --target PATH` commands; these alone are\n  not a deletion authorization or complete reference proof. `\u002Ftmp` ownership is\n  unknown unless the current policy names `tmp_owner`; record explicit release\n  receipts with the owning lane. Changing a plan's owner or release field cannot enable it;\n  even a released lease still requires a separate implementation of complete\n  reference checking before any `\u002Ftmp` apply can be supported. Tracked follow-up:\n  `host-hygiene-tmp-references`.\n- Inaccessible volume sizes need the exact diagnostic `sudo du` inspection;\n  unreadable tools\u002Fdata retain that class rather than skipping checks.\n\nLive concurrent ownership changes cannot be atomically excluded by CLI snapshots.\nCoordinate the exact resource lease with its owner before apply; a failed delete\nstops the run without retrying or widening scope. Earlier success\u002Ffailure receipts\nare emitted even when apply exits nonzero; a failed Nix batch explicitly reports\nthat partial deletion is possible. If SSH or the outer timeout loses contact,\nremote work may still be running: inspect it and reconcile the exact saved list\nbefore retrying. Never force Docker removal. After three\nfailed attempts at a repair approach, report the evidence and escalate. Preserve\nthe plan and partial results, measure again, and review a new plan for any retry.\n\nRun `scripts\u002Ftest-host-hygiene.sh` for the disposable SSH\u002FCLI fixture suite and\nguard-removal mutations. The suite never connects to a real host.\n","Follow orgs `doctrine\u002FDOCTRINE.md`, “Host resources and hygiene”. The owning\nlane inventories and reviews; a **fresh Utility agent** receives exactly that\nreviewed list and applies it. Authorization for hygiene does not authorize\nkilling peers, deleting worktrees or peer caches, or changing shared tags.\n\nUse `scripts\u002Fhygiene.sh plan HOST PLAN.json --policy POLICY.json`. Linux,\nPython 3, SSH BatchMode, Docker and Nix tools are expected; nothing uses sudo.\nThe per-host policy must specify an absolute `projects_root` on the remote host.\nDo not substitute the local user's home-expanded root: another host or SSH user\nmay keep projects elsewhere. The [shared configuration](..\u002F_shared\u002Fconfiguration.md)\nCLI can supply a candidate only after you confirm that path on the target host.\nThe script runs over SSH without installing or creating remote helper files.\nUse a private local state directory, e.g. `~\u002F.local\u002Fstate\u002Fhygiene\u002FYYYYMMDD\u002FHOST\u002F`.\nPlan files are created exclusively with mode 0600. The printed SHA256 binds the\nwhole reviewed list; neither edit the list nor substitute a new inventory after\nreview. Before deletion, inspect each full ID\u002Fname\u002Fpath, size in bytes, ownership,\nreason, eligibility, and every diagnostic. Missing sizes are never deletable.\nBuildx versions may render human sizes: `size_display` preserves them,\n`size_exact=false` marks the `bytes` value as an estimate rather than an exact\nbyte measurement. Unsupported size formats disable cache eligibility. `\u002Ftmp`\nuses allocated inode blocks for nondirectories. Directories use GNU `du` in\nbatches of at most 256 exact paths, with NUL input\u002Foutput and a shared 60s\nmeasurement budget. Symlinks are not followed; path aliases, mounted roots or directories\ncontaining mounts are retained without traversal. Directory sizes count every\nhardlink to avoid cross-operand deduplication and are marked `size_exact=false`:\nthey can exceed an independent per-root deduplicated measurement and do not\npredict freed bytes. Missing\u002Funsupported output, stderr, nonzero exit, timeout,\nor changed directory\u002Fmount identity discards every size in that affected batch;\nall entries remain in the plan with unknown sizes and diagnostics. Failed batches\ninclude output byte counts and at most 512-byte excerpts per stream; subprocess\ncapture itself is not a streaming memory bound. CLI snapshots\ncannot exclude a concurrent mount or inode change atomically.\nMeasure `df -h` before and after separately and retain the plan and apply output.\n\nA small owner-maintained policy looks like this (replace full IDs):\n\n```json\n{\n  \"projects_root\": \"\u002Fabsolute\u002Fpath\u002Fto\u002Fprojects\",\n  \"records\": [\"\u002Fabsolute\u002Fpath\u002Fto\u002Fcurrent-owner-and-paused-lane-record.md\"],\n  \"protected\": [\"peer-service\"],\n  \"tmp_owner\": \"peer-lane\",\n  \"builder\": \"default\",\n  \"claims\": [\n    {\"kind\": \"container\", \"id\": \"FULL_64_HEX_ID\", \"owner\": \"my-lane\", \"idle\": true},\n    {\"kind\": \"image\", \"id\": \"sha256:FULL_64_HEX_ID\", \"owner\": \"my-lane\", \"idle\": true,\n     \"released_tags\": [\"my-owned-repo:old-tag\"]},\n    {\"kind\": \"volume\", \"id\": \"my-owned-compose-volume\", \"owner\": \"my-lane\", \"idle\": true},\n    {\"kind\": \"cache\", \"id\": \"FULL_BUILDKIT_CACHE_ID\", \"owner\": \"my-lane\", \"idle\": true}\n  ]\n}\n```\n\n`records` is the explicit, nonempty list of required local ownership\u002Freference\ndocuments. Include current ownership and pause records for every project on\nthe target host; confirm their paths with the owning lanes.\nThe owning lane confirms completeness; missing\u002Funreadable records disable all\neligibility. Incomplete remote project-root\u002Fcompose scans disable Docker eligibility\nwhile independent Nix dead-graph\u002Froot checks remain usable.\nReference scans have a 60s budget, reject symlink\u002Fnonregular compose files,\nand fail closed on symlink directories instead of silently missing external\nworktree references. Review external references explicitly and declare complete\ncoverage in the policy; do not blindly follow arbitrary links or loops.\nAn incomplete scan retains Docker\u002Fcache entries. Compose files below\n`projects_root`, including paused worktrees, conservatively protect references,\nexcept intentionally excluded `.git`, `.bare`, `node_modules` and `.venv` trees.\nThe owner must provide relevant compose\u002Flane references from excluded trees or\nother unscanned coverage in `records`; exclusions do not establish absence of\nreferences. Add other project roots' compose\u002Flane information as local\nrecords; the owning lane must confirm the declared root covers its host's worktrees.\n`protected` adds exact names, full IDs, tags or paths from peer leases and records.\nPaused is not abandoned: their images, volumes, unrooted store paths and caches\nmust be named in records or `protected`; do not grant their ownership to yourself.\n`claims` are affirmative owner-confirmed idle assertions, not inferred from age.\nConflicting claims for one resource, including a peer's non-idle claim, retain it;\none lane's idle claim cannot borrow another lane's tag-release authorization.\nFor images, `released_tags` is an explicit release of every current tag by its\nowner; shared tags may not be released by this lane.\n\nGive the fresh Utility agent the reviewed plan, digest, policy and owning lane:\n\n```sh\nscripts\u002Fhygiene.sh apply HOST PLAN.json --policy POLICY.json \\\n  --owner my-lane --reviewed-sha256 REVIEWED_DIGEST > APPLY.json\n```\n\nApply reloads current policy\u002Frecords and inventories the same host\u002Fuser, then\nmatches only saved eligible entries against fresh IDs, inspection fingerprints,\nages, references and ownership. It checks the exact Docker\u002Fbackend identity again\nbefore each deletion and rechecks the Nix dead graph before one exact-list daemon\ndelete batch (referrers and references can be deleted together).\nDocker removes exited containers older than 24h without force\u002Fvolume removal;\nuntagged unused images; tagged unused images older than 14d with explicit tag\nrelease; and unreferenced compose volumes whose size and mount identity can be\nread. No broad prune runs. Nix inventories roots and the dead graph with sizes;\nthe exact reviewed dead-path subset is passed as one batch to `nix-store --delete`, which\nrechecks liveness through the daemon. No roots\u002Fprofiles\u002Fgenerations are removed,\nand newly eligible paths are never added to the reviewed list. Store mtimes are\nnormalized and do not prove installation age; generation-age cleanup is deferred\nunder follow-up `host-hygiene-profile-age`. Keep paused\nunrooted closures protected explicitly; the daemon protects rooted closures.\n\nConcrete retained classes:\n\n- Build cache unused >72h is inventoried as NDJSON on the explicit `builder`.\n  Only immutable, private, reclaimable, owner-confirmed records qualify. Apply\n  pins the builder\u002Fbackend identity and uses ANDed `id=FULL_ID`, `until=72h`,\n  `private=true`, `inuse=false`, and `immutable=true` filters, then verifies the\n  selected ID is absent. Mutable\u002Fshared\u002Fpaused\u002Funowned entries are retained;\n  unsupported JSON or filter options fail closed without a broad-prune fallback.\n  See [Docker Buildx prune](https:\u002F\u002Fdocs.docker.com\u002Freference\u002Fcli\u002Fdocker\u002Fbuildx\u002Fprune\u002F)\n  and [disk usage](https:\u002F\u002Fdocs.docker.com\u002Freference\u002Fcli\u002Fdocker\u002Fbuildx\u002Fdu\u002F).\n- `\u002Ftmp` entries older than 2d are enumerated with exact paths and sizes, but no\n  `\u002Ftmp` deletion is implemented: unprivileged access cannot prove complete\n  process, mount and socket references across peers. Diagnostics name exact\n  `sudo lsof +D PATH` and `sudo findmnt --target PATH` commands; these alone are\n  not a deletion authorization or complete reference proof. `\u002Ftmp` ownership is\n  unknown unless the current policy names `tmp_owner`; record explicit release\n  receipts with the owning lane. Changing a plan's owner or release field cannot enable it;\n  even a released lease still requires a separate implementation of complete\n  reference checking before any `\u002Ftmp` apply can be supported. Tracked follow-up:\n  `host-hygiene-tmp-references`.\n- Inaccessible volume sizes need the exact diagnostic `sudo du` inspection;\n  unreadable tools\u002Fdata retain that class rather than skipping checks.\n\nLive concurrent ownership changes cannot be atomically excluded by CLI snapshots.\nCoordinate the exact resource lease with its owner before apply; a failed delete\nstops the run without retrying or widening scope. Earlier success\u002Ffailure receipts\nare emitted even when apply exits nonzero; a failed Nix batch explicitly reports\nthat partial deletion is possible. If SSH or the outer timeout loses contact,\nremote work may still be running: inspect it and reconcile the exact saved list\nbefore retrying. Never force Docker removal. After three\nfailed attempts at a repair approach, report the evidence and escalate. Preserve\nthe plan and partial results, measure again, and review a new plan for any retry.\n\nRun `scripts\u002Ftest-host-hygiene.sh` for the disposable SSH\u002FCLI fixture suite and\nguard-removal mutations. The suite never connects to a real host.\n",{"name":5,"description":7},"master","f233dc9c278370625755a94f72c1ee3815214b89","skills\u002Fhost-hygiene","https:\u002F\u002Fgithub.com\u002Ftedks\u002Fagent-skills","https:\u002F\u002Fgithub.com\u002Ftedks","https:\u002F\u002Ftedks.com","2026-11-03 02:32:59",{"source":8,"skillName":5,"commands":32,"prompt":42,"zipAvailable":43},[33,36,39],{"tool":34,"command":35},"npx","npx skills add tedks\u002Fagent-skills --skill host-hygiene",{"tool":37,"command":38},"bunx","bunx skills add tedks\u002Fagent-skills --skill host-hygiene",{"tool":40,"command":41},"pnpm","pnpm dlx skills add tedks\u002Fagent-skills --skill host-hygiene","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Ftedks\u002Fagent-skills\u002Fblob\u002Fmaster\u002Fskills\u002Fhost-hygiene\u002FSKILL.md",true,[45,51,55],{"path":46,"resourceType":47,"resourceTypeDesc":48,"sizeBytes":49,"sourceHtmlUrl":50},"scripts\u002Fhygiene.py","SCRIPT","脚本",29013,"https:\u002F\u002Fgithub.com\u002Ftedks\u002Fagent-skills\u002Fblob\u002Fmaster\u002Fskills\u002Fhost-hygiene\u002Fscripts\u002Fhygiene.py",{"path":52,"resourceType":47,"resourceTypeDesc":48,"sizeBytes":53,"sourceHtmlUrl":54},"scripts\u002Fhygiene.sh",144,"https:\u002F\u002Fgithub.com\u002Ftedks\u002Fagent-skills\u002Fblob\u002Fmaster\u002Fskills\u002Fhost-hygiene\u002Fscripts\u002Fhygiene.sh",{"path":56,"resourceType":47,"resourceTypeDesc":48,"sizeBytes":57,"sourceHtmlUrl":58},"scripts\u002Ftest-host-hygiene.sh",31374,"https:\u002F\u002Fgithub.com\u002Ftedks\u002Fagent-skills\u002Fblob\u002Fmaster\u002Fskills\u002Fhost-hygiene\u002Fscripts\u002Ftest-host-hygiene.sh",[60,67,74,81,88,95],{"skillNo":61,"name":62,"nameZh":6,"description":63,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":64,"categories":65,"keywords":66,"publishedAt":19},"01M7EME6XZ35HJRKWWKGM3TCF4","where","Use before any find, rg, grep or Glob over a repo or a directory you do not already have an exact path for. Prints a small map of the current project (root, worktrees, depth-2 tree with sizes, the repo's own \"Project structure\" notes) so you search the worktree you own with a path instead of enumerating ~, ~\u002FProjects, \u002F, \u002Fnix or \u002Ftmp.","https:\u002F\u002Fgithub.com\u002Ftedks\u002Fagent-skills\u002Fblob\u002Fmaster\u002Fskills\u002Fwhere\u002FSKILL.md",[],[],{"skillNo":68,"name":69,"nameZh":6,"description":70,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":71,"categories":72,"keywords":73,"publishedAt":19},"01M7EME20D4MY2J5NMD9XMH4HP","usage","Use when a CTO or lead asks how much its lanes are spending, which session is burning quota, whether a lane tripped the token tripwire, or where the provider quota meters stand. Reads the installed token-usage counts (turns\u002Fh, cached tokens\u002Fh, average context per turn for each session; per-project\u002Fmodel totals; quota meters) and the tripwire's open incidents. Counts only; never message content. Does not install anything or change thresholds.","https:\u002F\u002Fgithub.com\u002Ftedks\u002Fagent-skills\u002Fblob\u002Fmaster\u002Fskills\u002Fusage\u002FSKILL.md",[],[],{"skillNo":75,"name":76,"nameZh":6,"description":77,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":78,"categories":79,"keywords":80,"publishedAt":19},"01M7EMDZKTR8RD7QM6TB5SYWRZ","stacked-prs","Detailed workflow for managing stacked PRs. Use when rebasing a PR stack, resolving merge conflicts in a stack, merging stacked PRs, or when user mentions \"stack\" and \"PR\" together.","https:\u002F\u002Fgithub.com\u002Ftedks\u002Fagent-skills\u002Fblob\u002Fmaster\u002Fskills\u002Fstacked-prs\u002FSKILL.md",[],[],{"skillNo":82,"name":83,"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},"01M7EMDVP1P5RNE1S6VYGWPNFW","spawn-agent","Spawn AI agents in steerable tmux windows for parallel or long-running work, including transcript-preserving Codex session forks with step-scoped compaction and structured final-recap handoffs. Use to spawn interactive agents, steer running tmux agents, or fork a Codex session for a separate step.","https:\u002F\u002Fgithub.com\u002Ftedks\u002Fagent-skills\u002Fblob\u002Fmaster\u002Fskills\u002Fspawn-agent\u002FSKILL.md",[],[],{"skillNo":89,"name":90,"nameZh":6,"description":91,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":92,"categories":93,"keywords":94,"publishedAt":19},"01M7EMDR15E49C5D52XRJ52PK7","retire-agent","Preserve a private, resumable agent handoff and verify a polite exit, or transfer the handoff to a freshly spawned successor.","https:\u002F\u002Fgithub.com\u002Ftedks\u002Fagent-skills\u002Fblob\u002Fmaster\u002Fskills\u002Fretire-agent\u002FSKILL.md",[],[],{"skillNo":96,"name":97,"nameZh":6,"description":98,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":99,"categories":100,"keywords":101,"publishedAt":19},"01M7EMDJYJR0EF8QG0XN0GCV7T","lean-automation","Diagnose and prevent CI\u002Fdeploy latency caused by automation inheriting a human's environment. Use when CI or a deploy feels slow, when auditing a repo's pipeline, when adding a dependency to a shared dev environment (flake devShell, base image, requirements file), or when writing a guard that is supposed to keep an environment lean.","https:\u002F\u002Fgithub.com\u002Ftedks\u002Fagent-skills\u002Fblob\u002Fmaster\u002Fskills\u002Flean-automation\u002FSKILL.md",[],[]]