安全

Skill Sync

试用

Audit and converge local skills across Codex, Claude, OpenClaw, OpenCode, workspace skills, and shared agent libraries; compute a hygiene score, classify sha...

它能做什么

Audit and converge local skills across Codex, Claude, OpenClaw, OpenCode, workspace skills, and shared agent libraries; compute a hygiene score, classify sha...

技能文档

Skill Sync

When To Use

Use this skill when the user wants to:

  • scan local skills across Codex, Claude, OpenClaw, OpenCode, workspace ./skills, and shared roots
  • see which skills are shared, duplicate, compatible, specific, or mixed
  • choose one canonical source and replace duplicate copies with symlinks
  • preview or apply convergence onto one preferred root such as ~/.agents/skills
  • export a portable layout manifest and import the same topology on another machine
  • restore a previous dedupe or import run

Do not use this skill when the task is only about one host and no cross-host comparison or symlink management is needed.

Fast Path

Scan the machine and get the hygiene score plus recommended actions:

python3 scripts/skill_sync.py

List only shared and host-specific skills:

python3 scripts/skill_sync.py --status shared,specific --list-names

Inspect one conflicting portable skill:

python3 scripts/skill_sync.py --diff rapid-ocr

Preview safe dedupe:

python3 scripts/skill_sync.py --dedupe --strategy strict

Preview a single-root convergence plan:

python3 scripts/skill_sync.py --adopt-root agents

Apply convergence with backups:

python3 scripts/skill_sync.py --adopt-root agents --apply

Restore the latest run:

python3 scripts/skill_sync.py --restore latest
python3 scripts/skill_sync.py --restore latest --apply

Workflow

  1. Scan first. Run python3 scripts/skill_sync.py and read the hygiene score and recommended actions.
  2. Review risky groups before mutation. Use --status compatible --list-names, --status mixed --list-names, and --diff .
  3. Preview convergence. Use --dedupe --strategy strict for identical groups, or --adopt-root to converge around one root.
  4. Apply only when the plan looks right. Add --apply to execute symlink creation or replacement.
  5. Export or import machine layouts when needed. Use --export-manifest and --import-manifest for cross-machine reuse.
  6. Restore from backup if needed. Use --restore and add --apply to roll back.

Safety Rules

  • The script always scans and reports before mutation.
  • It only auto-links portable directory-based skills that contain SKILL.md.
  • It never overwrites an existing destination without first moving it into ~/.skill-sync/backups//originals/....
  • --strategy strict only dedupes identical portable skills.
  • --strategy prefer-latest and --strategy trust-high may select the newest portable copy when content differs.
  • --restore latest --apply replays the last backup manifest in reverse.

Roots Scanned

  • /skills
  • ~/.codex/skills
  • ~/.agents/skills
  • ~/.claude/skills
  • ~/.claude/skills/anthropic-skills/skills
  • ~/.config/opencode/skills
  • ~/.openclaw/skills
  • ~/.openclaw/extensions/*/skills

Status Model

  • shared: the same real directory is visible from multiple hosts, usually through symlinks
  • duplicate: multiple hosts have the same portable skill content, but not via the same real path yet
  • compatible: multiple hosts have portable SKILL.md skills with the same name, but different content
  • specific: the skill appears on only one host
  • mixed: the same skill name exists on multiple hosts, but with different formats or incompatible content

If you need the detection details or compatibility notes, read references/compatibility.md.

相关技能

Use when users ask to sync, restore, apply, back up, push, or roll back local agent instructions, rules, or skills for Codex, Claude Code, OpenCode, Grok Build, CodeBuddy Code, WorkBuddy, or another Agent Skills-compatible platform.

Drive the Skills Manager CLI (`skm`) to initialize the hub, adopt unmanaged skills, list/enable/disable skills per AI tool, and doctor/fix symlink sync. Use whenever the user or an agent needs to manage skills from a terminal, SSH session, CI job, or headless machine; when a skill is missing in Claude Code, Codex, Cursor, Gemini, OpenCode, or any other supported tool; when asked to run skm / Skills Manager CLI; or when setting up Skills Manager without the GUI. Do not use for `npx skills` / skills.sh search (different CLI), for writing SKILL.md content (skill-creator), or for marketplace browse/install (not in skm).

Claude Code skill management. Topics — writer (create), lint (validate + fix frontmatter), merge (combine related), dedup (find duplicates), convert (agent → skill), architecture (multi-topic structure), upgrade (enhance + add topics), route (topic placement), trigger (declare + auto-register hooks), find (discover via npx skills CLI), graph (extract depends-on + topic body Skill calls into Edge Table + Mermaid + dispatched d3 force-directed render), language (per-skill language consistency enforcement), portability (public skill cross-ref + vendor isolation), publish-scope (published skill scope review before extending), invoke-discipline (slash command → Skill tool call, multi-topic Read, post-decision auto-invoke, interactive script, vendor dispatch). Use when: "skill writer", "skill lint", "skill merge", "skill dedup", "create skill", "frontmatter fix", "multi-topic skill", "convert agent", "skill upgrade", "add topic", "topic route", "trigger compile", "hook auto register", "find

24 次安装