MogScore Skill

Agent skills for running the official MogScore free AI Face Rating preview from https://mogscore.ai/mog-rating?mode=upload.

The workflow must use a visible, headed browser. Do not run it in headless Chrome; local face scan can fail with scanner_unavailable when browser graphics capabilities are missing.

Install With npx

Install the skill into a local agent runtime:

npx mogscore-skill install codex
npx mogscore-skill install claude-code
npx mogscore-skill install hermes
npx mogscore-skill install openclaw

Options:

npx mogscore-skill install codex --dir /custom/skills/mog-rating-preview
npx mogscore-skill install codex --force
npx mogscore-skill install codex --dry-run --json
npx mogscore-skill install codex --skip-deps

Default destinations:

  • Codex: ${CODEX_HOME:-~/.codex}/skills/mog-rating-preview
  • Claude Code: ~/.claude/skills/mog-rating-preview
  • Hermes: ~/.agents/skills/mog-rating-preview
  • OpenClaw local: ./skills/mog-rating-preview

The installer copies the selected skill and runs:

npm install --omit=dev --ignore-scripts --fund=false --audit=false

inside the installed skill folder. The package uses playwright-core, so this does not download Playwright's bundled Chromium. Pass --skip-deps when the host runtime already provides dependencies.

Install From ClawHub

After the OpenClaw skill is published:

clawhub install @roooo-798/mog-rating-preview

Publish manually:

clawhub login
clawhub skill publish openclaw/mog-rating-preview \
  --slug mog-rating-preview \
  --name "Mog Rating Preview" \
  --owner roooo-798

The repository also includes .github/workflows/clawhub-publish.yml, which uses OpenClaw's reusable workflow. Configure the CLAWHUB_TOKEN repository secret before enabling real publishes.

Install With skills.sh

The repository exposes a platform-neutral skill at skills/mog-rating-preview for skills.sh:

npx skills add roooo-798/mogscore-skill --skill mog-rating-preview

To preview without installing:

npx skills add roooo-798/mogscore-skill --list

Submit To SkillsLLM

SkillsLLM accepts public GitHub repositories through its submit page. The repository includes SKILL.md files in subdirectories and uses the MIT-0 license. SkillsLLM currently asks submitters to sign in with GitHub and expects public repos with at least 100 stars.

Runtime Use

Each variant exposes the same workflow contract:

  1. Open https://mogscore.ai/mog-rating?mode=upload.
  2. Attach a local image to the input labeled Selfie image.
  3. Click Get Score or Get my score.
  4. Capture the POST /api/reports JSON response.

If the host browser cannot upload files or capture response bodies, run the fallback script from the installed skill folder:

node scripts/run-mog-rating-preview.mjs \
  --browser-executable "/Applications/Google Chrome.app/Contents/MacOS/Google Chrome" \
  --headed \
  --image /absolute/path/to/photo.jpg

Development

npm install
npm test
npm run check
npm pack --dry-run
npm publish --dry-run --access public

Push source:

gh repo create roooo-798/mogscore-skill --public --source . --remote origin --push

End-to-end smoke tests require a user-supplied, non-sensitive, authorized 18+ portrait photo and a headed browser.