[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7AM8WV3KWJYJZ6SD4ZAMDTM":3,"skill-same:01M7AM8WV3KWJYJZ6SD4ZAMDTM":46},{"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":30,"updatedAt":31,"install":32,"resources":45},"01M7AM8WV3KWJYJZ6SD4ZAMDTM","yard-control-plane",null,"Audit and change the user's agent setup across Claude Code, Codex, and Cursor. Use when asked what is loaded, what is eating the context window, why a hook or MCP server is not working, or to enable\u002Fdisable a skill, plugin, or MCP server.","yannelli\u002Fyard","yannelli","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F59575788?v=4","01M7AM8JWB152290AM2XJ35MSV","Ryan Yannelli","01M7AM8JX0A9S735JR6B5T9WJ5",0,"MIT","https:\u002F\u002Fgithub.com\u002Fyannelli\u002Fyard\u002Fblob\u002Fmain\u002Fplugins\u002Fyard\u002Fskills\u002Fyard-control-plane\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: yard-control-plane\ndescription: Audit and change the user's agent setup across Claude Code, Codex, and Cursor. Use when asked what is loaded, what is eating the context window, why a hook or MCP server is not working, or to enable\u002Fdisable a skill, plugin, or MCP server.\n---\n\n# Yard\n\nYard reads the user's real configuration for all three clients and can change it. It is not limited\nto this repository.\n\n## When to reach for it\n\n- \"why is my context window so full\" → `env_context`\n- \"what skills\u002Fplugins\u002FMCP servers do I have\" → `env_inventory`\n- \"my hook isn't firing\" \u002F \"this MCP server won't start\" → `env_doctor`\n- \"turn off X\" → `env_set_skill`, `env_set_plugin`, `env_set_mcp`\n\n## Reading\n\n`env_inventory` returns counts plus a bounded list. Filter with `client` and `kind` rather than\npulling everything — there are often several hundred skills.\n\n`env_context` prices what enters the context window every turn: skill listing lines, MCP tool\nschemas, subagent descriptions, slash commands, and memory files. Numbers are estimates and are\nlabelled as such.\n\n`env_doctor` finds hooks pointing at deleted scripts, MCP servers that fail to start, skills with\nmissing or truncated frontmatter, plugins enabled but not installed, and duplicate skill names.\n\n`probe: true` on either one starts the user's configured stdio MCP servers to read their real tool\nlists. That executes their commands, so only pass it when the user asked for real numbers, and say\nthat is what you are doing.\n\n## Changing\n\nEvery mutating tool takes `dryRun`. Use it first when the change is not trivially reversible, show\nthe user what would change, then apply.\n\nWrites are atomic and backed up to `~\u002F.yard\u002Fbackups`, and each result names the file that changed.\n\nSkill visibility is the main lever for context, and only Claude Code has all four values:\n\n- `on` — name and description in the listing\n- `name-only` — name only, description dropped\n- `user-invocable-only` — invisible to the model, still available from `\u002F`\n- `off` — gone\n\n`user-invocable-only` is usually the right suggestion for a skill the user triggers by hand: it\ncosts nothing per turn and stays one keystroke away. Prefer it over `off` unless they want it gone.\n\nIf a name exists in more than one client the call fails and names the candidates. Pass `client`.\n\nCodex owns `~\u002F.codex\u002Fconfig.toml`, so MCP changes there are refused with the `codex mcp` command to\nrun instead — rewriting that file would destroy the user's comments and ordering.\n\n## Authoring\n\n`catalog_search` and `catalog_get` browse this marketplace's own plugins. `plugin_adapt` fills in\nthe files a Claude-only skill or plugin is missing for Codex and Cursor; it only writes what is\nabsent.\n\nOne thing worth knowing when editing plugins: plugin hooks are a single PascalCase `hooks\u002Fhooks.json`\nthat all three clients read. The camelCase `version: 1` form belongs to the user's own\n`.cursor\u002Fhooks.json` and is not a plugin file. Cursor also reads plugin MCP servers from\n`.cursor-plugin\u002Fplugin.json`, not from `.mcp.json`.\n","# Yard\n\nYard reads the user's real configuration for all three clients and can change it. It is not limited\nto this repository.\n\n## When to reach for it\n\n- \"why is my context window so full\" → `env_context`\n- \"what skills\u002Fplugins\u002FMCP servers do I have\" → `env_inventory`\n- \"my hook isn't firing\" \u002F \"this MCP server won't start\" → `env_doctor`\n- \"turn off X\" → `env_set_skill`, `env_set_plugin`, `env_set_mcp`\n\n## Reading\n\n`env_inventory` returns counts plus a bounded list. Filter with `client` and `kind` rather than\npulling everything — there are often several hundred skills.\n\n`env_context` prices what enters the context window every turn: skill listing lines, MCP tool\nschemas, subagent descriptions, slash commands, and memory files. Numbers are estimates and are\nlabelled as such.\n\n`env_doctor` finds hooks pointing at deleted scripts, MCP servers that fail to start, skills with\nmissing or truncated frontmatter, plugins enabled but not installed, and duplicate skill names.\n\n`probe: true` on either one starts the user's configured stdio MCP servers to read their real tool\nlists. That executes their commands, so only pass it when the user asked for real numbers, and say\nthat is what you are doing.\n\n## Changing\n\nEvery mutating tool takes `dryRun`. Use it first when the change is not trivially reversible, show\nthe user what would change, then apply.\n\nWrites are atomic and backed up to `~\u002F.yard\u002Fbackups`, and each result names the file that changed.\n\nSkill visibility is the main lever for context, and only Claude Code has all four values:\n\n- `on` — name and description in the listing\n- `name-only` — name only, description dropped\n- `user-invocable-only` — invisible to the model, still available from `\u002F`\n- `off` — gone\n\n`user-invocable-only` is usually the right suggestion for a skill the user triggers by hand: it\ncosts nothing per turn and stays one keystroke away. Prefer it over `off` unless they want it gone.\n\nIf a name exists in more than one client the call fails and names the candidates. Pass `client`.\n\nCodex owns `~\u002F.codex\u002Fconfig.toml`, so MCP changes there are refused with the `codex mcp` command to\nrun instead — rewriting that file would destroy the user's comments and ordering.\n\n## Authoring\n\n`catalog_search` and `catalog_get` browse this marketplace's own plugins. `plugin_adapt` fills in\nthe files a Claude-only skill or plugin is missing for Codex and Cursor; it only writes what is\nabsent.\n\nOne thing worth knowing when editing plugins: plugin hooks are a single PascalCase `hooks\u002Fhooks.json`\nthat all three clients read. The camelCase `version: 1` form belongs to the user's own\n`.cursor\u002Fhooks.json` and is not a plugin file. Cursor also reads plugin MCP servers from\n`.cursor-plugin\u002Fplugin.json`, not from `.mcp.json`.\n",{"name":5,"description":7},"main","c01ef2efd30476fa7619d5bb211eca6899cca76e","plugins\u002Fyard\u002Fskills\u002Fyard-control-plane","https:\u002F\u002Fgithub.com\u002Fyannelli\u002Fyard","https:\u002F\u002Fgithub.com\u002Fyannelli","https:\u002F\u002Fryanyannelli.com","#Engineer #HealthTech #T1D","2026-08-17 05:37:14",{"source":8,"skillName":5,"commands":33,"prompt":43,"zipAvailable":44},[34,37,40],{"tool":35,"command":36},"npx","npx skills add yannelli\u002Fyard --skill yard-control-plane",{"tool":38,"command":39},"bunx","bunx skills add yannelli\u002Fyard --skill yard-control-plane",{"tool":41,"command":42},"pnpm","pnpm dlx skills add yannelli\u002Fyard --skill yard-control-plane","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fyannelli\u002Fyard\u002Fblob\u002Fmain\u002Fplugins\u002Fyard\u002Fskills\u002Fyard-control-plane\u002FSKILL.md",true,[],[47,54,61],{"skillNo":48,"name":49,"nameZh":6,"description":50,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":51,"categories":52,"keywords":53,"publishedAt":19},"01M7AM8ZMBEKNXC0BFV4GW2ZY4","PLUGIN_NAME","PLUGIN_DESCRIPTION","https:\u002F\u002Fgithub.com\u002Fyannelli\u002Fyard\u002Fblob\u002Fmain\u002Ftemplates\u002Fplugin\u002Fskills\u002FPLUGIN_NAME\u002FSKILL.md",[],[],{"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},"01M7AM8V3B0S9XBJ6JTKXAQR69","review","Review selected or recent code for bugs, security, and performance. Use when asked to review a diff, PR, or the current file.","https:\u002F\u002Fgithub.com\u002Fyannelli\u002Fyard\u002Fblob\u002Fmain\u002Fplugins\u002Freview\u002Fskills\u002Freview\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":15,"licenseFileUrl":6,"sourceHtmlUrl":65,"categories":66,"keywords":67,"publishedAt":19},"01M7AM8RDC202VW4J6QWXDRYRV","hello","Greet the user and confirm the yannelli-skills marketplace is installed. Use when asked to say hello, verify the marketplace, or smoke-test a plugin install.","https:\u002F\u002Fgithub.com\u002Fyannelli\u002Fyard\u002Fblob\u002Fmain\u002Fplugins\u002Fhello\u002Fskills\u002Fhello\u002FSKILL.md",[],[]]