skillpack-harvest

Copy mature host skills into gbrain with privacy lint and dry-run preview.

1|Updated May 9, 2026
One-click install
npx skills add https://github.com/weiping/gbrain-cn --skill skillpack-harvest
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: skillpack-harvest
Source: https://github.com/weiping/gbrain-cn/tree/main/skills/skillpack-harvest
Command: npx skills add https://github.com/weiping/gbrain-cn --skill skillpack-harvest

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill prevents successful but private or fork-specific “host” skills from being incorrectly upstreamed into gbrain by enforcing an editorial genericization workflow plus a privacy lint and verification gates.

Core Features & Use Cases

  • Upstream skill lifting (host → gbrain): Copies a mature host skill into gbrain’s skills/<slug>/ directory and updates openclaw.plugin.json so other gbrain clients can scaffold it.
  • Editorial genericization and trigger scrubbing: Rewrites fork-specific names, entities, channels, and routing examples so triggers and evaluation fixtures generalize safely.
  • Safety-checked publishing workflow: Performs a dry-run preview, runs privacy lint (unless explicitly bypassed with documentation), rolls back on lint failure, and validates with conformance tests after harvest.

Quick Start

Use the skillpack harvest workflow to copy a mature host skill into gbrain by running a dry-run preview first, then performing the real harvest once the editorial genericization checklist and privacy lint pass for the chosen slug.

Frequently Asked Questions about skillpack-harvest

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I migrate host skills into a gbrain bundle without leaking private patterns?

Migrating host skills into a gbrain bundle safely requires an editorial genericization workflow with a privacy lint. This process rewrites fork-specific triggers and enforces a rollback on lint failure to prevent leaking private patterns.

What is the safest way to package host skills for cross-client use?

The safest way to package host skills for cross-client use involves a dry-run file set preview followed by privacy linting and YAML frontmatter validation. This ensures triggers and routing-eval fixtures generalize correctly for other clients.

Can I preview file changes before lifting a host skill into gbrain?

Yes, you can preview file changes before lifting a host skill into gbrain by running a dry-run file set preview. This allows you to verify the copied triggers and paired source files before updating the openclaw.plugin.json manifest.

Do I need to update the openclaw plugin manifest when harvesting skills?

Yes, updating the openclaw.plugin.json manifest is required when harvesting skills. This manifest update registers the newly generalized skill in the gbrain bundle so other clients can successfully scaffold it.

Why does skill harvesting require editorial genericization before publishing?

Skill harvesting requires editorial genericization before publishing to rewrite fork-specific entities, channels, and routing evaluation examples. This step ensures that triggers and evaluation fixtures generalize safely across different gbrain clients.