global-skill

النجوم٢

التفرعات٠

آخر تحديث٢٢ يوليو ٢٠٢٦ في ٠١:٤٥

Read, create, update, or delete globally available agent skills provisioned by this Nix configuration. Use for any CRUD operation on shared skills for Claude Code, Codex, OpenCode, or Copilot CLI.

التثبيت

التثبيت باستخدام Codex أو Claude انسخ هذا Prompt والصقه في Codex أو Claude أو مساعد آخر ليراجع صفحة Skill ويثبّتها لك.

المصدر

George-Miao

George-Miao/nix-config

فتح مستودع GitHub عرض مستودعات المنشئ

تنزيل

المهن ذات الصلةSOC

استنادا إلى تصنيف SOC المهني

المهن الحاسوبية الأخرىمهن الحاسوب والرياضيات·SOC 15-1299

مستكشف الملفات

2 ملفات

SKILL.md

readonly

name	global-skill
description	Read, create, update, or delete globally available agent skills provisioned by this Nix configuration. Use for any CRUD operation on shared skills for Claude Code, Codex, OpenCode, or Copilot CLI.

Manage Global Skills

Locate this Nix configuration at $HOME/.nix-config, read AGENTS.md, and use skills/ as the source of truth. Never inspect secrets/ or edit generated copies under ~/.claude, ~/.codex, ~/.config/opencode, or ~/.copilot.
Determine whether the request is to read, create, update, or delete a skill. Inspect existing skills and repository references before changing anything. For a read, list, inspect, or explain the source skill without modifying it.
Use a concise skill name containing only lowercase letters, digits, and single hyphens. Do not overwrite or delete an existing skill unless the user requested that operation. For deletion, resolve the exact skills/<name> directory and remove references that would otherwise break evaluation.
For creation or update, use the current agent's skill-authoring workflow or scaffolder when available. Otherwise edit skills/<name>/SKILL.md directly and add only resources the workflow needs.
Put only name and description in YAML frontmatter for compatibility across all four agents. Make the description state what the skill does and when it should trigger. Write the body as concise imperative instructions.
Keep scripts in scripts/, detailed material in references/, and output templates in assets/. Test added or changed scripts. Do not add auxiliary files such as a README or changelog.
Preserve unrelated work. When updating or deleting, remove obsolete skill files and metadata only when they belong to the requested skill.
Validate every created or updated skill with the available skill validator. Then run nix fmt and git diff --check; run broader repository checks when warranted. Do not activate or deploy merely to validate.
Report the CRUD operation, affected skill paths, and validation results. Ask whether to reapply system activation with rb; after activation, the repository-managed skills become globally available to all four configured agents.

المزيد من هذا المستودع

نفس المستودع

commit-changes

George-Miao/nix-config

Split pending changes in a user-specified Git repository or the current workspace into meaningful, contained local Conventional Commits without pushing. Use when the user asks to commit local work, organize a worktree into commits, or create commits without publishing them.

2026-07-222

nix-package

George-Miao/nix-config

Read, add, update, configure, or remove globally provisioned Nix packages in this Nix configuration. Use for CRUD operations on Home Manager, NixOS, or nix-darwin packages across Linux, Darwin, or both.

2026-07-222

publish-nix-config

George-Miao/nix-config

Split pending changes in $HOME/.nix-config into meaningful, contained Conventional Commits and push them to the current branch. Use when the user asks to commit, publish, or push changes in the Nix configuration repository.

2026-07-222

Manage Global Skills

Locate this Nix configuration at $HOME/.nix-config, read AGENTS.md, and use skills/ as the source of truth. Never inspect secrets/ or edit generated copies under ~/.claude, ~/.codex, ~/.config/opencode, or ~/.copilot.

Determine whether the request is to read, create, update, or delete a skill. Inspect existing skills and repository references before changing anything. For a read, list, inspect, or explain the source skill without modifying it.

Use a concise skill name containing only lowercase letters, digits, and single hyphens. Do not overwrite or delete an existing skill unless the user requested that operation. For deletion, resolve the exact skills/<name> directory and remove references that would otherwise break evaluation.

For creation or update, use the current agent's skill-authoring workflow or scaffolder when available. Otherwise edit skills/<name>/SKILL.md directly and add only resources the workflow needs.

Put only name and description in YAML frontmatter for compatibility across all four agents. Make the description state what the skill does and when it should trigger. Write the body as concise imperative instructions.

Keep scripts in scripts/, detailed material in references/, and output templates in assets/. Test added or changed scripts. Do not add auxiliary files such as a README or changelog.

Preserve unrelated work. When updating or deleting, remove obsolete skill files and metadata only when they belong to the requested skill.

Validate every created or updated skill with the available skill validator. Then run nix fmt and git diff --check; run broader repository checks when warranted. Do not activate or deploy merely to validate.

Report the CRUD operation, affected skill paths, and validation results. Ask whether to reapply system activation with rb; after activation, the repository-managed skills become globally available to all four configured agents.