[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M77H2CQK4VHNB2P7HMSAVDWC":3,"skill-same:01M77H2CQK4VHNB2P7HMSAVDWC":53},{"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":31,"commitSha":32,"skillDirectory":33,"forksCount":34,"repositoryHtmlUrl":35,"repositoryHomepageUrl":36,"ownerHtmlUrl":37,"ownerBlogUrl":38,"ownerDescription":6,"updatedAt":39,"install":40,"resources":52},"01M77H2CQK4VHNB2P7HMSAVDWC","memory-compact",null,"Compact and structure a Claude Code auto-memory store so MEMORY.md stays under the harness load cap (first 200 lines \u002F 25KB load each session — content past that SILENTLY never loads) while every memory stays reachable. Splits the index into two tiers — hot MEMORY.md (laws\u002Fpreferences\u002Flive-keys\u002Fin-flight\u002Fopen-research) + cold MEMORY_archive.md (shipped\u002Ffixed\u002Fforensic\u002Fdated, recalled on demand) — and proves \"done\" with an integrity witness (check_memory.py) that re-derives both caps + the both-files bijection from disk, never from narration. Use when the operator says.","anthony-chaudhary\u002Ffak","anthony-chaudhary","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F18080164?v=4","01M77GSKCAEKV3JS07J40B94DY","Anthony Chaudhary","01M77GSKD26Z7F4FQQD4VA30VX",48,"Apache-2.0","https:\u002F\u002Fgithub.com\u002Fanthony-chaudhary\u002Ffak\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fmemory-compact\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: memory-compact\ndescription: Compact and structure a Claude Code auto-memory store so MEMORY.md stays under the harness load cap (first 200 lines \u002F 25KB load each session — content past that SILENTLY never loads) while every memory stays reachable. Splits the index into two tiers — hot MEMORY.md (laws\u002Fpreferences\u002Flive-keys\u002Fin-flight\u002Fopen-research) + cold MEMORY_archive.md (shipped\u002Ffixed\u002Fforensic\u002Fdated, recalled on demand) — and proves \"done\" with an integrity witness (check_memory.py) that re-derives both caps + the both-files bijection from disk, never from narration. Use when the operator says.\ndisable-model-invocation: false\nuser-invocable: true\nallowed-tools: Read, Edit, Write, Grep, Glob, Bash, AskUserQuestion\nargument-hint: \"[--dir PATH] [--check]   (--check = witness only, no edits)\"\noutput_root: none\nmetadata:\n  generated-by: fak project-assets sync\n  canonical: ..\u002F..\u002F..\u002F.claude\u002Fskills\u002Fmemory-compact\u002FSKILL.md\n---\n\n# \u002Fmemory-compact — keep the auto-memory index under the harness cap, provably\n\nThe recurring problem this skill packages: **`MEMORY.md` grows past the harness\nload cap and silently truncates, so the bottom of the index becomes invisible to\nevery future session — yet a session still reports \"memory looks fine.\"** That is\nthe narration-vs-ground-truth gap, inside the memory store itself. The fix is\nstructural (two tiers by value) and the proof is a witness (re-derive the\ninvariants from disk, let the exit code be the verdict).\n\n## Finding the store\n\nThe Claude Code auto-memory dir for a project is under the Claude Code config\nroot, keyed by the encoded working-tree path:\n\n```\n~\u002F.claude\u002Fprojects\u002F\u003Cencoded-cwd>\u002Fmemory\n```\n\nThe encoded form replaces `\u002F` and `:` with `-` (e.g. a tree at `\u002Fwork\u002Fmyproj` →\n`-work-myproj`). On Windows the root is `%USERPROFILE%\\.claude\\projects\\...`. Pass\n`--dir` to point the witness at the exact path. If the store is small (well under\ncap) this skill is a no-op — it is here so that when the index grows past the cap,\nthe fix is a one-command pass, not a guess.\n\n## The one ground-truth fact everything rests on\n\nThe Claude Code harness injects **only the first 200 lines OR 25KB of `MEMORY.md`**\n(whichever comes first) at session start. Content past that threshold **does not\nload** — not truncated-with-a-warning in a way the model can act on, just gone.\nA sibling file (`MEMORY_archive.md`, a topic `.md`) is reachable **only** if a\nsession sees a link to it in the loaded part of `MEMORY.md` and chooses to `Read`\nit. So:\n\n- shrinking `MEMORY.md` under the cap is the *only* thing that fixes what loads;\n- moving an entry to a sibling file keeps it reachable **iff** the index still\n  links it (one hop) — otherwise you have hidden it, which is worse than truncation.\n\nThis is verified harness behavior, not a guess. If a future harness changes the\ncap, change `--max-lines` \u002F `--max-bytes` in the witness and this doc.\n\n## The two-tier structure (split by value)\n\nA flat index has no notion of *value* — a refuted 3-week-old spike costs the same\nalways-loaded line as a load-bearing law. Split by a cheap-frequent-vs-rare logic:\n\n- **`MEMORY.md` = HOT** (always loaded): working preferences, architecture\u002Fdesign\n  laws, live keys, **in-flight work**, **open research**, live reference cards.\n- **`MEMORY_archive.md` = COLD** (recalled on demand): **shipped features** (done +\n  in the code), **fixed bugs**, **dated process residue** (handoffs\u002Fnotes-passes\u002F\n  release logs), **forensic-once audits**, **superseded design notes**.\n\n**Default going forward** (also stated in the `MEMORY.md` header so every session\ninherits it): a NEW memory that is shipped\u002Ffixed\u002Fforensic\u002Fdated goes to\n`MEMORY_archive.md`; only hot-tier kinds get a line in `MEMORY.md`. When a hot\nentry's work ships or its conclusion is absorbed into a law, MOVE it cold.\n\n## The invariant that must never break\n\n**The index↔disk bijection spans BOTH files**: every topic `.md` on disk is\nreferenced **exactly once** across `MEMORY.md` + `MEMORY_archive.md` — no orphan\nfiles (on disk, unreferenced → invisible forever), no orphan links (referenced,\nabsent → a broken promise), no duplicate references. The witness checks this.\n\n## Procedure\n\n1. **Witness first (baseline).** Run the script — it tells you exactly what is\n   broken and by how much. This is also the whole skill if invoked with `--check`.\n\n   ```bash\n   python .claude\u002Fskills\u002Fmemory-compact\u002Fcheck_memory.py --dir \"\u003Cmemory dir>\"\n   ```\n   Exit 0 = clean; non-zero = a hard invariant (cap or bijection) failed.\n\n2. **If over the LINE cap** — the structural lever. Trim any multi-line header\n   prose to a few lines (the how-to belongs in the topic file, not the\n   always-loaded index), then move\u002Fdelete cold entries until under 200 lines.\n\n3. **If over the BYTE cap** — compress hooks. The detail already lives in the\n   topic file, so an index line needs only Title + the one load-bearing verdict +\n   the primary handle (SHA \u002F `wf_` id \u002F doc name). Don't shred the dense research\n   lines to a lossy stub — ~200–250 chars that keeps the SHA + conclusion beats a\n   100-char stub that loses the recall key.\n\n4. **Tiering move (the main lever):** to move an entry cold, copy its index line\n   **byte-identical** into the right section of `MEMORY_archive.md`, then delete it\n   from `MEMORY.md`. Tiering is *where it lives*, NOT a re-summarize (re-summarize\n   is step 3, a separate concern). Keep the `MEMORY_archive.md`\n   pointer in the `MEMORY.md` header so the cold tier stays one hop away.\n\n5. **Deletion needs sign-off.** Hard-deleting an entry (index line **and** topic\n   `.md` from disk) is irreversible — only do it with the operator's explicit OK,\n   and present the list first (use `AskUserQuestion`). Delete only: a refuted spike\n   whose idea is fully captured in a surviving law\u002Fentry, an exact duplicate, or a\n   superseded \"in-flight\" dup of a \"shipped\" entry. **MERGE ≠ DELETE** for a file\n   with a refuted *conclusion* but a unique *idea* — keep it, add a ⚠ caveat.\n\n6. **Witness last (the verdict).** Re-run step 1. Exit 0 is the only \"done.\" If a\n   prior session said \"fixed,\" do not believe it — the witness is the read-back for\n   a store that lives outside any git repo.\n\n## Mechanics that matter (if the store is MULTI-SESSION-HOT)\n\n- If several concurrent sessions may write `MEMORY.md`, the byte count rises\n  **between your edits**. You cannot hold a fixed byte target by hand against live\n  writers — so **leave margin**: the witness gates at 25_000 bytes, ~600B under the\n  real 25_600 cap, precisely so concurrent growth doesn't immediately re-truncate.\n  Aim for a comfortable line margin too (~150 lines), not 199.\n- The index **self-heals**: another session adds its own file + index line. Do NOT\n  race it by adding an index line for a file you didn't create — you collide with\n  their pending edit (a duplicate). Only fix an orphan that has sat un-indexed\n  across multiple witness runs.\n- Edit by **fresh-read → targeted Edit**; an external touch between Read and Edit\n  trips the \"modified since read\" guard. Re-read the region if an Edit fails.\n- **Dangling `[[wiki-links]]`** to deleted slugs are TOLERATED (the memory spec\n  treats an unresolved `[[name]]` as a write-later marker, not an error), so the\n  witness reports them as advisory and does not fail on them unless `--strict`.\n\n## The witness (`check_memory.py`)\n\nPure stdlib, read-only, ships beside this file. Checks (1) the 200-line \u002F 25KB\ncap on `MEMORY.md`, (2) the both-files bijection, (3) dangling wiki-links\n(advisory). `--json` for machine output, `--strict` to also fail on dangling\nlinks, `--max-lines`\u002F`--max-bytes` if the harness cap ever changes. Wire it into a\nStop hook to get a verdict automatically, or run it by hand after any pass.\n\n## Caveats \u002F honest limits\n\n- The memory dir lives **outside any git repo**, so nothing here is git-committed;\n  the files on disk are the only durability.\n- The cap figures (200 \u002F 25KB) are the harness's as of 2026-06. Re-verify with the\n  claude-code-guide if behavior seems to differ, and update the witness defaults.\n- This skill operates on memory files only — it never touches any code lane.\n","# \u002Fmemory-compact — keep the auto-memory index under the harness cap, provably\n\nThe recurring problem this skill packages: **`MEMORY.md` grows past the harness\nload cap and silently truncates, so the bottom of the index becomes invisible to\nevery future session — yet a session still reports \"memory looks fine.\"** That is\nthe narration-vs-ground-truth gap, inside the memory store itself. The fix is\nstructural (two tiers by value) and the proof is a witness (re-derive the\ninvariants from disk, let the exit code be the verdict).\n\n## Finding the store\n\nThe Claude Code auto-memory dir for a project is under the Claude Code config\nroot, keyed by the encoded working-tree path:\n\n```\n~\u002F.claude\u002Fprojects\u002F\u003Cencoded-cwd>\u002Fmemory\n```\n\nThe encoded form replaces `\u002F` and `:` with `-` (e.g. a tree at `\u002Fwork\u002Fmyproj` →\n`-work-myproj`). On Windows the root is `%USERPROFILE%\\.claude\\projects\\...`. Pass\n`--dir` to point the witness at the exact path. If the store is small (well under\ncap) this skill is a no-op — it is here so that when the index grows past the cap,\nthe fix is a one-command pass, not a guess.\n\n## The one ground-truth fact everything rests on\n\nThe Claude Code harness injects **only the first 200 lines OR 25KB of `MEMORY.md`**\n(whichever comes first) at session start. Content past that threshold **does not\nload** — not truncated-with-a-warning in a way the model can act on, just gone.\nA sibling file (`MEMORY_archive.md`, a topic `.md`) is reachable **only** if a\nsession sees a link to it in the loaded part of `MEMORY.md` and chooses to `Read`\nit. So:\n\n- shrinking `MEMORY.md` under the cap is the *only* thing that fixes what loads;\n- moving an entry to a sibling file keeps it reachable **iff** the index still\n  links it (one hop) — otherwise you have hidden it, which is worse than truncation.\n\nThis is verified harness behavior, not a guess. If a future harness changes the\ncap, change `--max-lines` \u002F `--max-bytes` in the witness and this doc.\n\n## The two-tier structure (split by value)\n\nA flat index has no notion of *value* — a refuted 3-week-old spike costs the same\nalways-loaded line as a load-bearing law. Split by a cheap-frequent-vs-rare logic:\n\n- **`MEMORY.md` = HOT** (always loaded): working preferences, architecture\u002Fdesign\n  laws, live keys, **in-flight work**, **open research**, live reference cards.\n- **`MEMORY_archive.md` = COLD** (recalled on demand): **shipped features** (done +\n  in the code), **fixed bugs**, **dated process residue** (handoffs\u002Fnotes-passes\u002F\n  release logs), **forensic-once audits**, **superseded design notes**.\n\n**Default going forward** (also stated in the `MEMORY.md` header so every session\ninherits it): a NEW memory that is shipped\u002Ffixed\u002Fforensic\u002Fdated goes to\n`MEMORY_archive.md`; only hot-tier kinds get a line in `MEMORY.md`. When a hot\nentry's work ships or its conclusion is absorbed into a law, MOVE it cold.\n\n## The invariant that must never break\n\n**The index↔disk bijection spans BOTH files**: every topic `.md` on disk is\nreferenced **exactly once** across `MEMORY.md` + `MEMORY_archive.md` — no orphan\nfiles (on disk, unreferenced → invisible forever), no orphan links (referenced,\nabsent → a broken promise), no duplicate references. The witness checks this.\n\n## Procedure\n\n1. **Witness first (baseline).** Run the script — it tells you exactly what is\n   broken and by how much. This is also the whole skill if invoked with `--check`.\n\n   ```bash\n   python .claude\u002Fskills\u002Fmemory-compact\u002Fcheck_memory.py --dir \"\u003Cmemory dir>\"\n   ```\n   Exit 0 = clean; non-zero = a hard invariant (cap or bijection) failed.\n\n2. **If over the LINE cap** — the structural lever. Trim any multi-line header\n   prose to a few lines (the how-to belongs in the topic file, not the\n   always-loaded index), then move\u002Fdelete cold entries until under 200 lines.\n\n3. **If over the BYTE cap** — compress hooks. The detail already lives in the\n   topic file, so an index line needs only Title + the one load-bearing verdict +\n   the primary handle (SHA \u002F `wf_` id \u002F doc name). Don't shred the dense research\n   lines to a lossy stub — ~200–250 chars that keeps the SHA + conclusion beats a\n   100-char stub that loses the recall key.\n\n4. **Tiering move (the main lever):** to move an entry cold, copy its index line\n   **byte-identical** into the right section of `MEMORY_archive.md`, then delete it\n   from `MEMORY.md`. Tiering is *where it lives*, NOT a re-summarize (re-summarize\n   is step 3, a separate concern). Keep the `MEMORY_archive.md`\n   pointer in the `MEMORY.md` header so the cold tier stays one hop away.\n\n5. **Deletion needs sign-off.** Hard-deleting an entry (index line **and** topic\n   `.md` from disk) is irreversible — only do it with the operator's explicit OK,\n   and present the list first (use `AskUserQuestion`). Delete only: a refuted spike\n   whose idea is fully captured in a surviving law\u002Fentry, an exact duplicate, or a\n   superseded \"in-flight\" dup of a \"shipped\" entry. **MERGE ≠ DELETE** for a file\n   with a refuted *conclusion* but a unique *idea* — keep it, add a ⚠ caveat.\n\n6. **Witness last (the verdict).** Re-run step 1. Exit 0 is the only \"done.\" If a\n   prior session said \"fixed,\" do not believe it — the witness is the read-back for\n   a store that lives outside any git repo.\n\n## Mechanics that matter (if the store is MULTI-SESSION-HOT)\n\n- If several concurrent sessions may write `MEMORY.md`, the byte count rises\n  **between your edits**. You cannot hold a fixed byte target by hand against live\n  writers — so **leave margin**: the witness gates at 25_000 bytes, ~600B under the\n  real 25_600 cap, precisely so concurrent growth doesn't immediately re-truncate.\n  Aim for a comfortable line margin too (~150 lines), not 199.\n- The index **self-heals**: another session adds its own file + index line. Do NOT\n  race it by adding an index line for a file you didn't create — you collide with\n  their pending edit (a duplicate). Only fix an orphan that has sat un-indexed\n  across multiple witness runs.\n- Edit by **fresh-read → targeted Edit**; an external touch between Read and Edit\n  trips the \"modified since read\" guard. Re-read the region if an Edit fails.\n- **Dangling `[[wiki-links]]`** to deleted slugs are TOLERATED (the memory spec\n  treats an unresolved `[[name]]` as a write-later marker, not an error), so the\n  witness reports them as advisory and does not fail on them unless `--strict`.\n\n## The witness (`check_memory.py`)\n\nPure stdlib, read-only, ships beside this file. Checks (1) the 200-line \u002F 25KB\ncap on `MEMORY.md`, (2) the both-files bijection, (3) dangling wiki-links\n(advisory). `--json` for machine output, `--strict` to also fail on dangling\nlinks, `--max-lines`\u002F`--max-bytes` if the harness cap ever changes. Wire it into a\nStop hook to get a verdict automatically, or run it by hand after any pass.\n\n## Caveats \u002F honest limits\n\n- The memory dir lives **outside any git repo**, so nothing here is git-committed;\n  the files on disk are the only durability.\n- The cap figures (200 \u002F 25KB) are the harness's as of 2026-06. Re-verify with the\n  claude-code-guide if behavior seems to differ, and update the witness defaults.\n- This skill operates on memory files only — it never touches any code lane.\n",{"name":5,"metadata":24,"description":7,"output_root":27,"allowed-tools":28,"argument-hint":29,"user-invocable":30,"disable-model-invocation":20},{"canonical":25,"generated-by":26},"..\u002F..\u002F..\u002F.claude\u002Fskills\u002Fmemory-compact\u002FSKILL.md","fak project-assets sync","none","Read, Edit, Write, Grep, Glob, Bash, AskUserQuestion","[--dir PATH] [--check]   (--check = witness only, no edits)",true,"main","f565b7b7db63e79649f3a02c08108132a86c5d05",".agents\u002Fskills\u002Fmemory-compact",16,"https:\u002F\u002Fgithub.com\u002Fanthony-chaudhary\u002Ffak","https:\u002F\u002Fanthony-chaudhary.github.io\u002Ffak\u002Fshowcase.html","https:\u002F\u002Fgithub.com\u002Fanthony-chaudhary","https:\u002F\u002Flinkedin.com\u002Fin\u002Fanthonychaudhary","2026-11-15 03:06:00",{"source":8,"skillName":5,"commands":41,"prompt":51,"zipAvailable":30},[42,45,48],{"tool":43,"command":44},"npx","npx skills add anthony-chaudhary\u002Ffak --skill memory-compact",{"tool":46,"command":47},"bunx","bunx skills add anthony-chaudhary\u002Ffak --skill memory-compact",{"tool":49,"command":50},"pnpm","pnpm dlx skills add anthony-chaudhary\u002Ffak --skill memory-compact","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fanthony-chaudhary\u002Ffak\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fmemory-compact\u002FSKILL.md",[],[54,61,70,77,84,89],{"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},"01M78V4059KRP2T88YR6Z05MJQ","serve-receipt","On-demand research protocol + deterministic receipt for serving a Hugging Face model on a concrete GPU target — fan out research lanes (model facts\u002Ffiles, engine support matrix, serving recipes, KV\u002Fstate memory math), compute the tiered feasibility verdict with `fak serve-receipt` (never narrated math), and emit a witnessed launch receipt. Use when asked for an \"initial config\", \"serving receipt\", or \"how do I serve \u003Cmodel> on \u003CGPUs>\".","https:\u002F\u002Fgithub.com\u002Fanthony-chaudhary\u002Ffak\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fserve-receipt\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":65,"licenseFileUrl":66,"sourceHtmlUrl":67,"categories":68,"keywords":69,"publishedAt":19},"01M77HRGKQV2CJ2K6ZFN4YQ32K","caveman","Ultra-compressed communication mode. Cuts output tokens 65% (measured) by speaking like caveman while keeping full technical accuracy. Supports intensity levels: lite, full (default), ultra, wenyan-lite, wenyan-full, wenyan-ultra. Use when user says \"caveman mode\", \"talk like caveman\", \"use caveman\", \"less tokens\", \"be brief\", or invokes \u002Fcaveman. Also auto-triggers when token efficiency is requested.","MIT","https:\u002F\u002Fgithub.com\u002Fanthony-chaudhary\u002Ffak\u002Fblob\u002Fmain\u002Finternal\u002Fsyspromptmmu\u002Fthird_party\u002Fcaveman\u002FLICENSE","https:\u002F\u002Fgithub.com\u002Fanthony-chaudhary\u002Ffak\u002Fblob\u002Fmain\u002Finternal\u002Fsyspromptmmu\u002Fthird_party\u002Fcaveman\u002FSKILL.md",[],[],{"skillNo":71,"name":72,"nameZh":6,"description":73,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":74,"categories":75,"keywords":76,"publishedAt":19},"01M77HRAY3Y3BE7APN5PVJF7NV","quality-report","Score the repo quality and emit a report card.","https:\u002F\u002Fgithub.com\u002Fanthony-chaudhary\u002Ffak\u002Fblob\u002Fmain\u002Finternal\u002Fctxplans\u002Ftestdata\u002Frepo\u002F.claude\u002Fskills\u002Fquality-report\u002FSKILL.md",[],[],{"skillNo":78,"name":79,"nameZh":6,"description":80,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":81,"categories":82,"keywords":83,"publishedAt":19},"01M77HR9807KHAX13DG9PW2K2Q","ctx-overlay","Compact the context window and prune the memory store when a session grows too long.","https:\u002F\u002Fgithub.com\u002Fanthony-chaudhary\u002Ffak\u002Fblob\u002Fmain\u002Finternal\u002Fctxplans\u002Ftestdata\u002Frepo\u002F.claude\u002Fskills\u002Fctx-overlay\u002FSKILL.md",[],[],{"skillNo":85,"name":72,"nameZh":6,"description":73,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":86,"categories":87,"keywords":88,"publishedAt":19},"01M77HR7XQ9X8F2CY9Z4F2ZEZ3","https:\u002F\u002Fgithub.com\u002Fanthony-chaudhary\u002Ffak\u002Fblob\u002Fmain\u002Finternal\u002Fctxplanlint\u002Ftestdata\u002Frepo\u002F.claude\u002Fskills\u002Fquality-report\u002FSKILL.md",[],[],{"skillNo":90,"name":79,"nameZh":6,"description":80,"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},"01M77HR5BVKCTR1WNJGRBSDNX7","https:\u002F\u002Fgithub.com\u002Fanthony-chaudhary\u002Ffak\u002Fblob\u002Fmain\u002Finternal\u002Fctxplanlint\u002Ftestdata\u002Frepo\u002F.claude\u002Fskills\u002Fctx-overlay\u002FSKILL.md",[],[]]