Skip to content
Back to skills

Ztx14 Agent Disagreement Arbitration

ASecurity

Apply a focused implementation and verification workflow for design and engineering arbitration in Isometric, React, responsive motion, or interactive UI work.

  • 14 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added November 7, 2026
designreactgitapi

Works with

  • api

Security analysis

A100/100

Scanned November 7, 2026

npx -y skills add Realjustawall/Perfect-AI --skill ztx14-agent-disagreement-arbitration --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Ztx14 Agent Disagreement Arbitration?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Ztx14 Agent Disagreement Arbitration
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/realjustawall-ztx14-agent-disagreement-arbitration/badge)](https://www.skillsdirectory.com/skills/realjustawall-ztx14-agent-disagreement-arbitration)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: ztx14-agent-disagreement-arbitration
description: Apply a focused implementation and verification workflow for design and engineering arbitration in Isometric, React, responsive motion, or interactive UI work.
---

# Design and Engineering Arbitration

## Activation
Use this skill when the task includes **design and engineering arbitration**, a related review, a bug report, or a regression test. Load the Isometric master skill first for site-specific installation and source conventions, then only other specialists that are directly relevant.

## Scope and implementation contract
Resolve conflicting recommendations by requirement priority, user value, accessibility, measured cost, and explicit owner decision.

Begin with task intent and current project evidence, load only the relevant small skill set, assign clear ownership, and give each agent a bounded deliverable. Never claim a test or a visual inspection occurred unless evidence exists.

## Execution sequence
1. **Inspect evidence.** Identify the real component, event/state path, installed source, app conventions, and available test commands. Record uncertain assumptions instead of silently filling gaps.
2. **Define the behavior.** Write down the expected initial, active, interrupted, cancelled, and resting states for this specific task. Name the property or state owner before editing.
3. **Implement narrowly.** Apply the smallest safe change that achieves the focus above. Reuse local utilities, tokens, and existing engine behavior before adding a dependency. Preserve unrelated files and styles.
4. **Cover input variants.** Exercise pointer/mouse where applicable, touch or coarse pointer, keyboard/focus, and an equivalent non-hover path. Check reduced-motion behavior and both light/dark themes when relevant.
5. **Measure and verify.** Run the applicable test, type check, and production build. Capture a deterministic screenshot or repeatable interaction trace for visible behavior. If a browser/device is unavailable, mark that check `NOT_TESTED`.

## Acceptance gates
- The behavior is implemented in real source code, not merely described in prose.
- Target selection and visible state match after repeated, rapid, interrupted, and cancelled interactions.
- The responsive layout remains usable at 320, 390, 768, and 1280 CSS px or the project’s actual breakpoints.
- Reduced motion keeps the same information and task outcome without unnecessary continuous movement.
- No new console error, hydration mismatch, stale event listener, idle animation loop, or accidental layout shift is introduced.
- The report names changed files, commands actually run, evidence captured, unresolved issues, and any limitation.

## Failure patterns to avoid
Do not load every skill, invent evidence, change shared files from parallel agents without ownership, or hide uncertainty.
- Avoid changing an unrelated theme, route, dependency, or interaction simply to make the local example easier.
- Do not claim success based solely on a source edit; distinguish code inspection, automated tests, and real-browser validation.

## References
- https://github.com/vercel-labs/agent-skills
- https://github.com/anthropics/skills
- https://github.com/addyosmani/web-quality-skills
- https://isometric.buildlab.in/llms.txt

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.