Everything needed to publish the Agent Skill Safety Kit on GitHub.

Copy/paste packet

# GitHub Publish Packet — Agent Skill Safety Kit

Use this packet to publish the Agent Skill Safety Kit as a public GitHub repo without rewriting copy.

## Repo metadata

- **Repo name:** `agent-skill-safety-kit`
- **Visibility:** public
- **Description:** Dependency-style safety review kit for AI-agent skills: permissions, memory, receipts, rollback, evals, vendor questions, and marketplace readiness.
- **Homepage:** `https://jarvislandingdeploy.vercel.app/agent-skill-safety-kit.html`
- **Topics:** `ai-agents`, `agent-safety`, `agent-skills`, `ai-security`, `tool-permissions`, `agent-memory`, `evals`, `openclaw`, `claude-code`, `marketplace-safety`

## Short launch copy

Agent skills are becoming software packages. Before you install one, list it in a marketplace, or delegate browser/SaaS/file/memory authority to it, review the permission boundary.

This repo gives you a free review kit: a 30-minute checklist, permission manifest template + JSON Schema, local validator, vendor questionnaire, marketplace checklist, sample mini-report, and submission assets.

Start here: `30-minute-review-kit.md`.

## First release title

`v0.1.0 — Agent Skill Safety Kit: permission manifests, validator, and vendor questionnaire`

## First release notes

This initial public release packages the Agent Skill Safety Kit for teams, marketplaces, and skill creators reviewing AI-agent skills before install, publication, sale, or delegation.

Included:

- 30-minute review checklist and verdict template
- Portable skill contract template
- Permission manifest template
- Permission manifest JSON Schema and example
- Browser-local manifest validator link
- Vendor/procurement questionnaire
- Marketplace safety checklist
- Redaction-friendly intake
- Sample mini-report
- Directory/community submission copy
- Share-ready visual preview assets

Safety boundary: do not paste secrets, credentials, customer data, private prompts, or proprietary internals into public issues or examples. Use fictional or redacted review cases.

## Copy/paste create flow

```bash
unzip agent-skill-safety-github-repo.zip
cd agent-skill-safety-kit
git init
git branch -M main
git add .
git commit -m "Initial Agent Skill Safety Kit"
gh repo create agent-skill-safety-kit --public \
  --description "Dependency-style safety review kit for AI-agent skills: permissions, memory, receipts, rollback, evals, vendor questions, and marketplace readiness." \
  --homepage "https://jarvislandingdeploy.vercel.app/agent-skill-safety-kit.html" \
  --source . \
  --remote origin \
  --push
```

Then add topics in GitHub UI or with `gh repo edit` when available:

```bash
gh repo edit --add-topic ai-agents --add-topic agent-safety --add-topic agent-skills --add-topic ai-security --add-topic tool-permissions --add-topic agent-memory --add-topic evals --add-topic openclaw --add-topic claude-code --add-topic marketplace-safety
```

## Manual publish checklist

- [ ] Confirm repo visibility is public.
- [ ] Confirm README renders correctly.
- [ ] Confirm download ZIP works from the landing page.
- [ ] Create release `v0.1.0` using the release notes above.
- [ ] Pin/link the live scorecard: https://jarvislandingdeploy.vercel.app/agent-skill-safety-scorecard.html
- [ ] Pin/link the local validator: https://jarvislandingdeploy.vercel.app/agent-skill-permission-manifest-validator.html
- [ ] Do not include private review cases or customer examples.