skills.rest
yanochka11
yanochka11COMMUNITY
Explore all skills by this author →
GitHub

build-review-interface

Build browser-based annotation interfaces for reviewing LLM traces with pass/fail labels.

Updated May 5, 2026
Search Tags:#human feedback#llm evaluation#playwright#annotation#data labeling#trace review
One-click install
npx skills add https://github.com/yanochka11/harness_bro --skill build-review-interface-yanochka11
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: build-review-interface
Source: https://github.com/yanochka11/harness_bro/tree/main/.claude/skills/curated/evals/build-review-interface
Command: npx skills add https://github.com/yanochka11/harness_bro --skill build-review-interface-yanochka11

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill helps teams quickly build custom annotation interfaces for reviewing LLM traces and collecting structured human feedback instead of relying on generic review workflows.

Core Features & Use Cases

  • Trace Review Interface: Builds browser-based annotation pages with Pass/Fail decisions, notes, navigation, and persistent labels for LLM evaluation data.
  • Domain-Specific Rendering: Adapts trace presentation for formats such as code, markdown, JSON, tables, tool calls, and conversational data while keeping important context accessible.
  • Evaluation Workflow Support: Adds reviewer productivity features such as keyboard shortcuts, filtering, reference panels, clustering, and Playwright-based validation for annotation tools.

Quick Start

Use the build-review-interface skill to create a browser annotation tool for my LLM traces with pass/fail labels, notes, and saved reviewer feedback.

Frequently Asked Questions about build-review-interface

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I build a custom annotation interface for reviewing LLM traces?▼

To build an LLM trace review interface, this Skill generates browser-based annotation pages equipped with Pass/Fail decisions, notes, and persistent labels to collect structured human feedback reliably.

Can I render different data formats like JSON and markdown in my trace review tool?▼

Yes, trace review interfaces can adapt domain-specific rendering for code, markdown, JSON, tables, tool calls, and conversational data, ensuring important context remains accessible during human review.

How do I add keyboard shortcuts and navigation controls to a data labeling workflow?▼

This Skill adds reviewer productivity features like keyboard shortcuts, filtering, and navigation controls directly into the annotation interface to streamline the data labeling workflow.

Does this annotation tool support Playwright validation for interface testing?▼

Yes, the Skill includes Playwright-based validation to verify interface reliability, ensuring your annotation tools for LLM evaluation function consistently during human review sessions.

What is the best way to collect structured human feedback on LLM evaluation data?▼

The best way to collect structured human feedback is using a custom annotation interface with persistent reviewer labels, notes, and trace rendering tailored to your specific LLM evaluation formats.

Are there limitations when building browser-based tools for LLM trace analysis?▼

Browser-based trace analysis tools require proper trace rendering, feedback persistence, and interface validation to function reliably, meaning complex formats may need domain-specific adaptation to keep context accessible.

Related Skills

Explore More Data & Analytics Skills→
kinnerkarmanishkinnerkarmanish

review-interface

Generate self-contained HTML annotation interfaces for reviewing agent outputs and traces.

Community
Intermediate
opendatahub-ioopendatahub-io
36

eval-review

Analyze evaluation results to identify alignment gaps between automated checks and human assessment.

Official
Advanced
nrl-ainrl-ai
11

login-flows

Implement reusable Playwright login patterns for username/password, OAuth/SSO, and MFA flows.

Official
Intermediate
hamelsmuhamelsmu
1.6k

build-review-interface

Generate a browser-based annotation interface for reviewing LLM traces with Pass/Fail feedback.

Community
Intermediate