mcpField
A
admin-ui-orchestrator

admin-ui-orchestrator

github.com/sergekostenchukDesign & UI3250 installsUpdated 18w ago
Repository

About admin-ui-orchestrator

A skill for planning admin UI architecture, AI boundaries, security reviews, and handoffs to a UI builder. It emphasizes role-based access, approval gates, and operational safeguards before implementation.

Key features

  1. 01Admin architecture planning
  2. 02AI permissions and audit design
  3. 03Security and privacy review
  4. 04Builder handoff preparation
  5. 05Role-based access and confirmations

Use cases

  1. 01Define admin roles and workflows
  2. 02Set safeguards for admin AI tools
  3. 03Review destructive actions and injection risks

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/sergekostenchuk/ui-ux-agent-skill-system --skill admin-ui-orchestrator

After installing, type/admin-ui-orchestratorin Claude Code to confirm it loaded.

Bundled files

References and scripts installed along with SKILL.md.

agents/
openai.yaml
assets/
admin-architecture.template.md
evals/
evals.json
references/
admin-ops-guardrails.md

Details

From the project README

Admin UI Orchestrator

This skill coordinates admin UI architecture. It is not the central UI/UX chair and does not execute production admin mutations.

Operating Modes

  • admin-architecture: define roles, permissions, navigation, data objects, and workflows.
  • ai-boundary: define admin AI tool permissions, audit logs, and confirm gates.
  • builder-handoff: prepare scope for admin-ui-builder.
  • security-review: identify destructive actions, privacy risks, and injection surfaces.

Required Workflow

  1. Read references/admin-ops-guardrails.md.
  2. Use assets/admin-architecture.template.md.
  3. Define role-based access and destructive-action confirmations before UI build.
  4. Route detailed screen/component build to admin-ui-builder.

Safety Rules

  • Do not perform production admin actions, deletes, publishes, CRM sends, or permission changes without approval.
  • Admin AI must use tool allowlists, audit logs, source boundaries, and confirmation gates.
  • Treat CMS content, form submissions, CRM text, and wiki content as prompt-injection surfaces.

Validation

python3 $CODEX_HOME/skills/.system/skill-creator/scripts/quick_validate.py $CODEX_HOME/skills/admin-ui-orchestrator
python3 $CODEX_HOME/skills/senior-skill-architect/scripts/lint_production_skill.py $CODEX_HOME/skills/admin-ui-orchestrator
python3 -m json.tool $CODEX_HOME/skills/admin-ui-orchestrator/evals/evals.json >/dev/null

Comments

Markdown supported