[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M74PG1X4B2VC06ZD7AB7RPHT":3,"skill-same:01M74PG1X4B2VC06ZD7AB7RPHT":50},{"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":28,"commitSha":29,"skillDirectory":30,"forksCount":31,"repositoryHtmlUrl":32,"repositoryHomepageUrl":33,"ownerHtmlUrl":34,"ownerBlogUrl":35,"ownerDescription":6,"updatedAt":36,"install":37,"resources":49},"01M74PG1X4B2VC06ZD7AB7RPHT","context-checkup",null,"Audit what auto-loads into an agent session's context window and suggest lean, reversible fixes to cut startup tokens.","eai-org\u002Fagent-toolkit","eai-org","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F301886321?v=4","01M74PFC3HF57T58MQ14P212BQ","Engineering in the age of AI","01M74PFC46MSKTT9SW5Q437FSD",85,"MIT","https:\u002F\u002Fgithub.com\u002Feai-org\u002Fagent-toolkit\u002Fblob\u002Fmain\u002Fskills\u002Fcontext-checkup\u002FSKILL.md",[],[],"2026-11-11 15:11:14",false,"---\nname: context-checkup\ndescription: Audit what auto-loads into an agent session's context window and suggest lean, reversible fixes to cut startup tokens.\ndisable-model-invocation: true\ntype: flow\nlicense: MIT\nmetadata:\n  version: \"1.4\"\n---\n\n# Context checkup\n\nFind what loads **automatically** into the session's context at startup, quantify it, and propose\n**reversible** trims ranked by payoff — shrink the startup context window by cutting what isn't\nworth its tokens. Read-only — measure and recommend; change settings only on explicit approval.\n\n## Mental model (spend attention top-down)\n\nCost hierarchy, biggest first:\n1. **MCP tool schemas** — large, but **deferred** by tool search. Only a problem if tool search is\n   OFF.\n2. **Always-on text** — every always-on instruction file (entry-file chain, `@imports`, unscoped\n   rules files) and **skill\u002Fagent\u002Fcommand descriptions**: every available one's description\n   loads every session, except skills flagged non-model-invocable (their description never\n   loads; a manual-only skill costs nothing until invoked).\n3. **MCP server instructions** — load per enabled server at startup; a chatty preamble can dwarf its\n   tool names.\n4. **MCP tool names** — ~10 tokens each. Cheap; 30 names ≈ 300 tokens isn't worth a recommendation.\n\nRules: **measure, never guess** — `wc`\u002Fcount each source, convert with tokens ≈ words × 1.3 for\nprose and characters \u002F 4 otherwise, and state that rule once in the table header. When the agent\nhas a built-in context breakdown command, ask the user to paste its output and use its numbers as\nthe baseline; estimate only what it does not break down. **Rank by tokens-saved × reversibility.**\nDistinguish **auto-loaded** from **lazy-linked** (nested AGENTS.md referenced from a parent are\n*not* loaded — don't flag them).\n\nTool-name counts, the skills\u002Fagents\u002Fcommands lists, and tool-search status come from the **session's\nsystem-reminders**, not disk — read them from your own context.\n\n## Checklist\n\nMechanics below are one agent's (settings keys, `~\u002F.claude` paths, which `CLAUDE_CONFIG_DIR`\nrelocates); on another agent, map each item to its equivalent or ask. `p`, `s` and `r` are the\nanchors defined under Useful probes.\n\n**0. Verify tool search is on first.** Evidence is the session itself: MCP tools listed as deferred\n(names only, searched on demand) means on, even once a search has pulled some schemas in; full\nschemas loaded from the start means off, and every \"names are cheap\" conclusion flips. Only then\nread the env var or setting (`ENABLE_TOOL_SEARCH`, incl. `auto:N` threshold modes) to explain why.\n\n**1. Always-on instruction files.** Measure the entry-file chain (`CLAUDE.md`, `.claude\u002FCLAUDE.md`,\n`CLAUDE.local.md`, or `AGENTS.md` and `.claude\u002FAGENTS.md` when read natively) in `p` and\n**every parent dir up to root**, the user-level entry file (`~\u002F.claude\u002FCLAUDE.md`), the\nmanaged-policy one (a file such as `\u002FLibrary\u002FApplication Support\u002FClaudeCode\u002FCLAUDE.md` or\n`\u002Fetc\u002Fclaude-code\u002FCLAUDE.md`, or the managed `claudeMd` setting), and every rules file,\nsubdirectories and symlinks included, in the project and user rules dirs (`$p\u002F.claude\u002Frules\u002F`,\n`~\u002F.claude\u002Frules\u002F`) that is not path-scoped (a `paths:` key in its `---` frontmatter; a body\n`paths:` line scopes nothing) — a project-rules symlink leading outside the project only once\nexternal imports are approved. Expand `@import` lines (recursive). `wc -w` each; an `AGENTS.md` or\nthe user-level file counts once, whether read natively, imported or met on the parent walk.\n\n**2. Descriptions (often the largest unexamined chunk).** Skim the in-session lists of **skills,\nsubagent types, and slash commands** — their descriptions are always loaded, except\nnon-model-invocable skills (see Mental model). Many rarely-used entries = standing cost. When the\nagent offers a skills list with token sizes or a visibility override, take the sizes from it; the\noverride (e.g. Claude Code's `skillOverrides`) hides a description without editing the skill.\n\n**3. MCP servers.** From `$s\u002F.mcp.json` (a `\u002Fcd` does not move it) + settings, list enabled\nservers. Count tool **names** per server; note any server with long **instructions**. Flag servers\nirrelevant to the user's work.\n\n**4. Enablement levers and hooks.** Read `~\u002F.claude\u002Fsettings.json`, `$p\u002F.claude\u002Fsettings.json`,\n`$r\u002F.claude\u002Fsettings.local.json` and, when `s` differs from `r`, the legacy\n`$s\u002F.claude\u002Fsettings.local.json`. Check `enableAllProjectMcpServers`, `enabledMcpjsonServers`,\n`enabledPlugins`. Precedence: managed > local > project > user. Each plugin can pull in MCP\nservers, skills, agents, hooks. Measure `SessionStart` hook output (from its script, or a run the\nuser approves) as a startup cost, and per-prompt `UserPromptSubmit` output as a cost per turn. Tool\nhooks' plain stdout stays out of the context, but their `additionalContext` enters it, a cost per\nmatching tool call.\n\n**5. Memory.** Lines of the memory index up to the agent's load cap (e.g. the first 200 lines or\n25KB of `MEMORY.md`, whichever comes first); topic files load on demand and are not a startup\ncost.\n\n## Output\n\nA short sizes table (source → ≈ tokens → auto-loaded? → relevant to user's work?), then a\nprioritized action list. For each action give the exact setting\u002Ffile edit, the rough savings, and\nhow to reverse it (prefer per-session toggles). Skip anything under ~a few hundred tokens unless\nasked to be exhaustive. Each action names the skill that performs it when available (look in the\nagent's skills and plugin dirs, as manual-only skills are missing from the in-session list), with\nthis skill's own namespace prefix if any: a fat entry file → `\u002Fagentify-project` (slimming into\ndocs); a fat memory index → `\u002Fmemory-doctor`. Then ask before editing.\n\n## Useful probes\n\n```bash\n# anchors — p: the session's primary working directory after any \u002Fcd ($PWD when the shell sits there)\n# s: the project root the session started in — p resolved below, right without a \u002Fcd; after one, or in a desktop-app worktree session (the main checkout root, which also serves its settings and hooks), set it by hand from the session's start context\n# r: where settings.local.json lives — the repo root, or the main checkout's root in a linked worktree; outside git, when that root is $HOME, on Windows, or when its .git\u002F.claude is not the user's: p\nc=${CLAUDE_CONFIG_DIR:-$HOME\u002F.claude}; p=$PWD; s=$(cd \"$p\" && pwd -P); g=$(git -C \"$s\" rev-parse --path-format=absolute --git-common-dir 2>\u002Fdev\u002Fnull); r=${g:+$(dirname \"$g\")}; { [ -n \"$r\" ] && [ \"$r\" != \"$HOME\" ]; } || r=$p\n# always-on instruction files: entry-file chain up to \u002F, user-level and managed entry files, unscoped rules\nd=$p; while :; do for f in AGENTS.md .claude\u002FAGENTS.md CLAUDE.md .claude\u002FCLAUDE.md CLAUDE.local.md; do [ \"$d\u002F$f\" = \"$c\u002FCLAUDE.md\" ] || { [ -f \"$d\u002F$f\" ] && wc -w \"$d\u002F$f\"; }; done; [ \"$d\" = \u002F ] && break; d=$(dirname \"$d\"); done\nwc -w \"$c\u002FCLAUDE.md\" \"\u002FLibrary\u002FApplication Support\u002FClaudeCode\u002FCLAUDE.md\" \u002Fetc\u002Fclaude-code\u002FCLAUDE.md 2>\u002Fdev\u002Fnull\nfind -L \"$p\u002F.claude\u002Frules\" \"$c\u002Frules\" -name '*.md' -exec sh -c 'for f; do awk '\"'\"'NR==1&&$0!=\"---\"{exit 1} NR>1&&$0==\"---\"{exit 1} NR>1&&\u002F^paths:\u002F{f=1;exit} END{exit !f}'\"'\"' \"$f\" || printf \"%s\\0\" \"$f\"; done' _ {} + 2>\u002Fdev\u002Fnull | xargs -0 -r wc -w\n# enabled MCP + settings\n{ [ -f \"$s\u002F.mcp.json\" ] && cat \"$s\u002F.mcp.json\"; } || echo \"no .mcp.json at $s\"; cat \"$p\u002F.claude\u002Fsettings.json\" \"$r\u002F.claude\u002Fsettings.local.json\" 2>\u002Fdev\u002Fnull; [ \"$s\" = \"$r\" ] || cat \"$s\u002F.claude\u002Fsettings.local.json\" 2>\u002Fdev\u002Fnull; cat \"$c\u002Fsettings.json\" 2>\u002Fdev\u002Fnull\n```\n","# Context checkup\n\nFind what loads **automatically** into the session's context at startup, quantify it, and propose\n**reversible** trims ranked by payoff — shrink the startup context window by cutting what isn't\nworth its tokens. Read-only — measure and recommend; change settings only on explicit approval.\n\n## Mental model (spend attention top-down)\n\nCost hierarchy, biggest first:\n1. **MCP tool schemas** — large, but **deferred** by tool search. Only a problem if tool search is\n   OFF.\n2. **Always-on text** — every always-on instruction file (entry-file chain, `@imports`, unscoped\n   rules files) and **skill\u002Fagent\u002Fcommand descriptions**: every available one's description\n   loads every session, except skills flagged non-model-invocable (their description never\n   loads; a manual-only skill costs nothing until invoked).\n3. **MCP server instructions** — load per enabled server at startup; a chatty preamble can dwarf its\n   tool names.\n4. **MCP tool names** — ~10 tokens each. Cheap; 30 names ≈ 300 tokens isn't worth a recommendation.\n\nRules: **measure, never guess** — `wc`\u002Fcount each source, convert with tokens ≈ words × 1.3 for\nprose and characters \u002F 4 otherwise, and state that rule once in the table header. When the agent\nhas a built-in context breakdown command, ask the user to paste its output and use its numbers as\nthe baseline; estimate only what it does not break down. **Rank by tokens-saved × reversibility.**\nDistinguish **auto-loaded** from **lazy-linked** (nested AGENTS.md referenced from a parent are\n*not* loaded — don't flag them).\n\nTool-name counts, the skills\u002Fagents\u002Fcommands lists, and tool-search status come from the **session's\nsystem-reminders**, not disk — read them from your own context.\n\n## Checklist\n\nMechanics below are one agent's (settings keys, `~\u002F.claude` paths, which `CLAUDE_CONFIG_DIR`\nrelocates); on another agent, map each item to its equivalent or ask. `p`, `s` and `r` are the\nanchors defined under Useful probes.\n\n**0. Verify tool search is on first.** Evidence is the session itself: MCP tools listed as deferred\n(names only, searched on demand) means on, even once a search has pulled some schemas in; full\nschemas loaded from the start means off, and every \"names are cheap\" conclusion flips. Only then\nread the env var or setting (`ENABLE_TOOL_SEARCH`, incl. `auto:N` threshold modes) to explain why.\n\n**1. Always-on instruction files.** Measure the entry-file chain (`CLAUDE.md`, `.claude\u002FCLAUDE.md`,\n`CLAUDE.local.md`, or `AGENTS.md` and `.claude\u002FAGENTS.md` when read natively) in `p` and\n**every parent dir up to root**, the user-level entry file (`~\u002F.claude\u002FCLAUDE.md`), the\nmanaged-policy one (a file such as `\u002FLibrary\u002FApplication Support\u002FClaudeCode\u002FCLAUDE.md` or\n`\u002Fetc\u002Fclaude-code\u002FCLAUDE.md`, or the managed `claudeMd` setting), and every rules file,\nsubdirectories and symlinks included, in the project and user rules dirs (`$p\u002F.claude\u002Frules\u002F`,\n`~\u002F.claude\u002Frules\u002F`) that is not path-scoped (a `paths:` key in its `---` frontmatter; a body\n`paths:` line scopes nothing) — a project-rules symlink leading outside the project only once\nexternal imports are approved. Expand `@import` lines (recursive). `wc -w` each; an `AGENTS.md` or\nthe user-level file counts once, whether read natively, imported or met on the parent walk.\n\n**2. Descriptions (often the largest unexamined chunk).** Skim the in-session lists of **skills,\nsubagent types, and slash commands** — their descriptions are always loaded, except\nnon-model-invocable skills (see Mental model). Many rarely-used entries = standing cost. When the\nagent offers a skills list with token sizes or a visibility override, take the sizes from it; the\noverride (e.g. Claude Code's `skillOverrides`) hides a description without editing the skill.\n\n**3. MCP servers.** From `$s\u002F.mcp.json` (a `\u002Fcd` does not move it) + settings, list enabled\nservers. Count tool **names** per server; note any server with long **instructions**. Flag servers\nirrelevant to the user's work.\n\n**4. Enablement levers and hooks.** Read `~\u002F.claude\u002Fsettings.json`, `$p\u002F.claude\u002Fsettings.json`,\n`$r\u002F.claude\u002Fsettings.local.json` and, when `s` differs from `r`, the legacy\n`$s\u002F.claude\u002Fsettings.local.json`. Check `enableAllProjectMcpServers`, `enabledMcpjsonServers`,\n`enabledPlugins`. Precedence: managed > local > project > user. Each plugin can pull in MCP\nservers, skills, agents, hooks. Measure `SessionStart` hook output (from its script, or a run the\nuser approves) as a startup cost, and per-prompt `UserPromptSubmit` output as a cost per turn. Tool\nhooks' plain stdout stays out of the context, but their `additionalContext` enters it, a cost per\nmatching tool call.\n\n**5. Memory.** Lines of the memory index up to the agent's load cap (e.g. the first 200 lines or\n25KB of `MEMORY.md`, whichever comes first); topic files load on demand and are not a startup\ncost.\n\n## Output\n\nA short sizes table (source → ≈ tokens → auto-loaded? → relevant to user's work?), then a\nprioritized action list. For each action give the exact setting\u002Ffile edit, the rough savings, and\nhow to reverse it (prefer per-session toggles). Skip anything under ~a few hundred tokens unless\nasked to be exhaustive. Each action names the skill that performs it when available (look in the\nagent's skills and plugin dirs, as manual-only skills are missing from the in-session list), with\nthis skill's own namespace prefix if any: a fat entry file → `\u002Fagentify-project` (slimming into\ndocs); a fat memory index → `\u002Fmemory-doctor`. Then ask before editing.\n\n## Useful probes\n\n```bash\n# anchors — p: the session's primary working directory after any \u002Fcd ($PWD when the shell sits there)\n# s: the project root the session started in — p resolved below, right without a \u002Fcd; after one, or in a desktop-app worktree session (the main checkout root, which also serves its settings and hooks), set it by hand from the session's start context\n# r: where settings.local.json lives — the repo root, or the main checkout's root in a linked worktree; outside git, when that root is $HOME, on Windows, or when its .git\u002F.claude is not the user's: p\nc=${CLAUDE_CONFIG_DIR:-$HOME\u002F.claude}; p=$PWD; s=$(cd \"$p\" && pwd -P); g=$(git -C \"$s\" rev-parse --path-format=absolute --git-common-dir 2>\u002Fdev\u002Fnull); r=${g:+$(dirname \"$g\")}; { [ -n \"$r\" ] && [ \"$r\" != \"$HOME\" ]; } || r=$p\n# always-on instruction files: entry-file chain up to \u002F, user-level and managed entry files, unscoped rules\nd=$p; while :; do for f in AGENTS.md .claude\u002FAGENTS.md CLAUDE.md .claude\u002FCLAUDE.md CLAUDE.local.md; do [ \"$d\u002F$f\" = \"$c\u002FCLAUDE.md\" ] || { [ -f \"$d\u002F$f\" ] && wc -w \"$d\u002F$f\"; }; done; [ \"$d\" = \u002F ] && break; d=$(dirname \"$d\"); done\nwc -w \"$c\u002FCLAUDE.md\" \"\u002FLibrary\u002FApplication Support\u002FClaudeCode\u002FCLAUDE.md\" \u002Fetc\u002Fclaude-code\u002FCLAUDE.md 2>\u002Fdev\u002Fnull\nfind -L \"$p\u002F.claude\u002Frules\" \"$c\u002Frules\" -name '*.md' -exec sh -c 'for f; do awk '\"'\"'NR==1&&$0!=\"---\"{exit 1} NR>1&&$0==\"---\"{exit 1} NR>1&&\u002F^paths:\u002F{f=1;exit} END{exit !f}'\"'\"' \"$f\" || printf \"%s\\0\" \"$f\"; done' _ {} + 2>\u002Fdev\u002Fnull | xargs -0 -r wc -w\n# enabled MCP + settings\n{ [ -f \"$s\u002F.mcp.json\" ] && cat \"$s\u002F.mcp.json\"; } || echo \"no .mcp.json at $s\"; cat \"$p\u002F.claude\u002Fsettings.json\" \"$r\u002F.claude\u002Fsettings.local.json\" 2>\u002Fdev\u002Fnull; [ \"$s\" = \"$r\" ] || cat \"$s\u002F.claude\u002Fsettings.local.json\" 2>\u002Fdev\u002Fnull; cat \"$c\u002Fsettings.json\" 2>\u002Fdev\u002Fnull\n```\n",{"name":5,"type":24,"license":15,"metadata":25,"description":7,"disable-model-invocation":27},"flow",{"version":26},"1.4",true,"main","12db59437c9d48b9bd62b18888e9974b17adafaf","skills\u002Fcontext-checkup",19,"https:\u002F\u002Fgithub.com\u002Feai-org\u002Fagent-toolkit","https:\u002F\u002Feai-org.github.io\u002Fagent-toolkit-docs","https:\u002F\u002Fgithub.com\u002Feai-org","https:\u002F\u002Fmedium.com\u002Fengineering-in-the-age-of-ai","2026-11-14 22:36:21",{"source":8,"skillName":5,"commands":38,"prompt":48,"zipAvailable":27},[39,42,45],{"tool":40,"command":41},"npx","npx skills add eai-org\u002Fagent-toolkit --skill context-checkup",{"tool":43,"command":44},"bunx","bunx skills add eai-org\u002Fagent-toolkit --skill context-checkup",{"tool":46,"command":47},"pnpm","pnpm dlx skills add eai-org\u002Fagent-toolkit --skill context-checkup","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Feai-org\u002Fagent-toolkit\u002Fblob\u002Fmain\u002Fskills\u002Fcontext-checkup\u002FSKILL.md",[],[51,57,64,71,78,85],{"skillNo":52,"name":53,"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":54,"categories":55,"keywords":56,"publishedAt":19},"01M74PKAVPTTY5PSSAWHCRVY9T","no-frontmatter","https:\u002F\u002Fgithub.com\u002Feai-org\u002Fagent-toolkit\u002Fblob\u002Fmain\u002Ftest\u002Fcompact-skill-creator\u002Ffixtures\u002Fno-frontmatter\u002FSKILL.md",[],[],{"skillNo":58,"name":59,"nameZh":6,"description":60,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":61,"categories":62,"keywords":63,"publishedAt":19},"01M74PK8J99W7V3669Y4JG2JKV","other-name","Fixture skill whose name differs from its folder.","https:\u002F\u002Fgithub.com\u002Feai-org\u002Fagent-toolkit\u002Fblob\u002Fmain\u002Ftest\u002Fcompact-skill-creator\u002Ffixtures\u002Fname-mismatch\u002FSKILL.md",[],[],{"skillNo":65,"name":66,"nameZh":6,"description":67,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":68,"categories":69,"keywords":70,"publishedAt":19},"01M74PK6T9DHVRCNHTEV2ZA4Y3","md-off","A fixture skill whose folder turns the line length off below a root that sets 99.","https:\u002F\u002Fgithub.com\u002Feai-org\u002Fagent-toolkit\u002Fblob\u002Fmain\u002Ftest\u002Fcompact-skill-creator\u002Ffixtures\u002Fmd-off\u002FSKILL.md",[],[],{"skillNo":72,"name":73,"nameZh":6,"description":74,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":75,"categories":76,"keywords":77,"publishedAt":19},"01M74PK4H4N3D290V4C08M4DGY","long-line","Fixture skill with one prose line of 100 characters, over the fixtures' limit of 99.","https:\u002F\u002Fgithub.com\u002Feai-org\u002Fagent-toolkit\u002Fblob\u002Fmain\u002Ftest\u002Fcompact-skill-creator\u002Ffixtures\u002Flong-line\u002FSKILL.md",[],[],{"skillNo":79,"name":80,"nameZh":6,"description":81,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":82,"categories":83,"keywords":84,"publishedAt":19},"01M74PK2DJ2MCR5KFPMYVQWNJG","flag-without-flow","Fixture skill with the flag and no type.","https:\u002F\u002Fgithub.com\u002Feai-org\u002Fagent-toolkit\u002Fblob\u002Fmain\u002Ftest\u002Fcompact-skill-creator\u002Ffixtures\u002Fflag-without-flow\u002FSKILL.md",[],[],{"skillNo":86,"name":87,"nameZh":6,"description":88,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":89,"categories":90,"keywords":91,"publishedAt":19},"01M74PK03XGEBQ986FDAE46WCM","double-backtick","A fixture skill with double-backtick code spans holding a link template and an overrun.","https:\u002F\u002Fgithub.com\u002Feai-org\u002Fagent-toolkit\u002Fblob\u002Fmain\u002Ftest\u002Fcompact-skill-creator\u002Ffixtures\u002Fdouble-backtick\u002FSKILL.md",[],[]]