agent-skill-migration
Skill1 starsChanged 53 days agoMITView on GitHub
Use when the user asks to package, migrate, back up, sync, install, or verify triggerability of local Codex, Claude Code, or agent skills/workflows across roots such as a user-selected skill workbench, .codex\skills, .agents\skills, and .claude\skills, especially for a new computer or cross-tool discoverability check. Also use to judge whether an external skill, plugin, or skills repo is worth adopting — 这个 skill 是什么/有什么用, 对我有没有帮助, 能给我带来什么帮助, 值不值得装, 看看这个 GitHub skills 仓库 — and to trace which local skills are self-authored versus installed.
1 other project uses a version of this.
Adopted by
Most recently taken up by
- memi-design/design-skillstheir own version, first seen Nov 8, 2026
Other versions1 project uses a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 1
- memi-design/design-skills★ 7same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository20 other files it gives its agents.
- llms.txtllms.txt
- agents-md-verification-mapSkill
- cocos-creator-mobile-gameSkill
- game-entity-movement-presentationSkill
- game-event-directorSkill
- game-ui-panelsSkill
- godot-narrative-workflowSkill
- narrative-game-prototype-plannerSkill
- obsidian-article-writebackSkill
- offline-ux-kitSkill
- product-image-prompt-iterationSkill
- product-teardownSkill
- reporting-disciplineSkill
- scene-anchor-systemSkill
- screen2layersSkill
- sol-luna-local-parallelSkill
- sprite-atlas-slicerSkill
- unfinished-work-auditSkill
- windows-ai-cli-recoverySkill
- xiaohongshu-copywriterSkill
The file
We split this skill into the rules it gives an agent
5 rules. About 0 are common, 1 is distinctive, and 4 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
4
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 26
Check a user-selected skill workbench as the source/workbench when present.
- Very specific: line 27
Check ~/.codex/skills, ~/.agents/skills, and ~/.claude/skills for discoverable copies.
- Distinctive: line 28
Do not claim Codex or Claude Code will trigger a workbench skill until its folder is present in the relevant user skill root.
- Very specific: line 31
Use PYTHONUTF8=1 on Windows before running the Codex quick_validate.py validator against UTF-8 skills with Chinese text.
- Very specific: line 36
Prefer reusing a user-provided migration packer if it exists and the task matches a full migration bundle.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Agent Skill Migration
Overview
Build a safe, copyable inventory or migration bundle for the user's local agent skills and workflow assets. Treat triggerability as an installation-root question, not a source-folder question.
For the full checklist, read references/migration-pack-checklist.md when the task involves creating a bundle, syncing skills into roots, or answering whether a source/workbench skill will trigger.
Decide The Scope
- For a triggerability question, inspect source folder plus installed roots and answer
will trigger / will not trigger / installed but description weak. - For a sync request, copy only the named skill folders into the requested visible roots and validate each copy.
- For a new-computer or backup request, create a single versioned folder and zip with manifests, install notes, and safety scan results.
- For a broad review, separate installed skills, workbench/source copies, plugin caches, workflow docs, and optional memory/context.
- For an adoption question about an external skill or repo(
对我有没有帮助), answer 装 / 不装 / 只借鉴,理由一句话。判据是补不补短板,不是这个 skill 写得好不好:先读它的真实SKILL.md(gh api repos/<owner>/<repo>/contents/...,不要只看 README 或凭名字猜),再对照已装 skill 找重叠,最后落到用户真实会重复遇到的场景。已有同类就说扩展哪个、别新装。 - For a provenance question(
哪些是我自己创建的), compare each local skill against GitHub by repo name and a distinctive SKILL.md sentence; 搜不到是强证据但不是铁证,说明置信度。注意多个本地 skill 顶部有用户自注的「全局协作约定」块,它的存在不代表整个 skill 是自创。
Required Checks
- Find live roots first.
- Check
a user-selected skill workbenchas the source/workbench when present. - Check
~/.codex/skills,~/.agents/skills, and~/.claude/skillsfor discoverable copies. - Do not claim Codex or Claude Code will trigger a workbench skill until its folder is present in the relevant user skill root.
- Check
- Validate skills after edits or sync.
- Use
PYTHONUTF8=1on Windows before running the Codexquick_validate.pyvalidator against UTF-8 skills with Chinese text. - Compare file counts or hashes between source and installed roots after copying.
- Re-read
SKILL.mdfrontmatter to ensure descriptions include the real projects and trigger words the user expects.
- Use
- Package safely.
- Prefer reusing
a user-provided migration packerif it exists and the task matches a full migration bundle. - Include manifests and routing inventory so the user can inspect what was packaged.
- Include redacted templates for config files that may contain tokens; do not copy raw auth settings.
- Scan both output folder and zip contents for secret-like strings before declaring the package ready.
- Prefer reusing
Output Contract
Report:
- roots inspected
- skills or workflow assets included/excluded
- whether each named skill is discoverable by Codex, agents, and Claude Code
- created folder/zip paths when packaging
- manifest counts
- validation and secret-scan results
- any roots or tools that were unavailable
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

