一键导入
contributing
Walk a workshop participant through contributing a skill or lab back to the Skills & Specs Lab via fork and pull request, including GitHub PAT setup
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
菜单
Walk a workshop participant through contributing a skill or lab back to the Skills & Specs Lab via fork and pull request, including GitHub PAT setup
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
| name | contributing |
| description | Walk a workshop participant through contributing a skill or lab back to the Skills & Specs Lab via fork and pull request, including GitHub PAT setup |
| state | active |
Use this skill when a participant wants to contribute a skill or a lab back to
the Skills & Specs Lab repository after the workshop. You guide them through the
full fork-and-pull-request flow. Read CONTRIBUTING.md
in the repository root first; it is the source of truth and this skill walks the
user through it.
This is a learning exercise. The goal is that the participant practices git, fork-and-pull, least-privilege credentials, and agent tooling, not contribution volume. Be encouraging and treat every step as teaching.
This is the first real least-privilege lesson, so do not rush it.
gridctl var set GITHUB_PERSONAL_ACCESS_TOKEN.Do not proceed until the token exists and is stored.
stack.yaml).upstream.git checkout -b skill/<name> or lab/<name>.skills-creator skill, then save the result under
showcase/skills/<name>/SKILL.md (the open floor), not skills/ (curated
core).showcase/labs/_template/ to showcase/labs/<name>/ and
fill it in.Remind the user that showcase content is instructions and configs other people will load and run, so it must be safe: no destructive commands, no secrets, no prompt-injection payloads. This is the supply-chain lesson made concrete.
main, filling in the PR template.At each checkpoint, confirm completion before advancing. End by summarizing what the user built, the PR URL, and one concrete next idea for their second contribution.
CONTRIBUTING.md; do not work around the policy.git CLI and the GitHub
web UI; the flow works either way.Snapshot a node's state before and after a change, then report exactly what changed and whether the fabric is still healthy
Query one fabric node and return its interface and BGP state as structured JSON
Summarize a single piece of fabric state in a structured way, instead of returning a raw tool dump (replace this with your own one-line, verb-first description that says when to use the skill)
Return a focused, structured health summary for one fabric node (interfaces and BGP), instead of a raw tool dump
Scaffold a new agentskills.io SKILL.md from a short description, with the right frontmatter, a focused scope, and a quality checklist