القائمة
Import a source folder into a knowledge-first skill by analyzing the folder with LLM assistance and generating SKILL.md plus skill.json.
التثبيت باستخدام Codex أو Claude انسخ هذا Prompt والصقه في Codex أو Claude أو مساعد آخر ليراجع صفحة Skill ويثبّتها لك.
استنادا إلى تصنيف SOC المهني
Build-focused skill for creating, updating, and installing standard Agent Skills.
Manage SOP templates, scheduled automation tasks, and run history through AI-callable tools.
提供独立的命令执行与 skill 脚本执行能力。
Query chat history by keywords and date ranges.
Provides interaction capabilities with the user.
Meta skill for AI self-evolution. Invoke when recording lessons learned and improving skill guidance from execution feedback.
| name | skill-importer |
| description | Import a source folder into a knowledge-first skill by analyzing the folder with LLM assistance and generating SKILL.md plus skill.json. |
| type | system |
| created_by | system |
| kind | system |
| allowed-tools | ["analyze_skill_source_folder","generate_skill_from_folder"] |
Use this skill when a user provides a folder and wants it converted into a reusable Cowork skill.
Use it to inspect a source folder, summarize its capabilities, suggest which lightweight tools are relevant, and generate skill metadata after user confirmation.
Use Skill Center for user-facing imports. It can now accept either a source folder or a ZIP package; ZIP packages are unpacked safely and then routed through the same Cowork skill adaptation path.
Do not use it for simple one-off file edits or when the user only wants an explanation of the folder contents.
analyze_skill_source_folder returns a JSON preview describing the inferred skill name, kind, candidate tool refs, references, risks, and generated metadata draft.
generate_skill_from_folder copies the source folder into ai_skills/<name> and writes SKILL.md plus skill.json centered on explanation, tool refs, and recommended workflow.
Skill Manager ZIP import supports flat-root packages and single-folder packages, reads the final skill name from skill.json or SKILL.md, rejects existing target names, and reloads skills after successful import.
Always preview before writing. Do not overwrite an existing generated skill unless the user has confirmed the update path.
When importing ZIP packages, never trust archive paths directly. The current Skill Manager validates that extracted paths stay inside the temporary extraction directory before adapting the package.
Folders often contain multiple scripts and partial documentation. Prefer documenting the smallest useful tools instead of turning the entire folder into one big executable workflow.
Keep the generated skill knowledge-first: document interfaces and caveats before introducing new executable wrappers.
