CommunityCoding & Developmentgithub.com

Rockiez/lit-panel

A blind-parallel literary review panel with 11 seats for Chinese memoir/narrative text, designed as a Claude Code plugin and Codex skill. Uses bands and verbatim evidence, never numeric scores.

What is lit-panel?

This skill, named Rockiez/lit-panel, provides a structured, blind-parallel literary review panel for evaluating Chinese memoir or narrative text. It leverages an 11-seat panel that assesses writing by assigning qualitative bands (e.g., superior, strong, adequate, weak) rather than numeric scores, ensuring feedback remains holistic and narrative-focused. Each panelist must cite verbatim evidence from the text to support their assessment, promoting objective and traceable critique. Designed primarily as a plugin for Claude Code and a skill for Codex, it integrates seamlessly into developer workflows for analyzing creative or personal writing in Chinese, making it ideal for authors, editors, or AI-assisted literary workshops.

Works with✓Claude Code✓Codex CLI~Cursor
npx skills add Rockiez/lit-panel

Installed? Explore more Coding & Development skills: steipete/bluebubbles, steipete/eightctl, steipete/blucli · View all 6 →

Ask in your favorite AI

Open a new chat with this agent skill pre-loaded.

Documentation

What does lit-panel do?

文学评审团 lit-panel — 11-seat blind-parallel literary review panel for Chinese memoir/narrative text. Claude Code plugin + Codex skill. Bands + verbatim evidence, never numeric scores.

Overview

Rockiez/lit-panel is a specialized AI agent skill for conducting blind-parallel literary reviews of Chinese memoir and narrative texts. It assembles 11 virtual panelists who independently evaluate the writing using qualitative bands (e.g., Superior, Strong, Adequate, Weak) and support their assessments with verbatim evidence from the text. No numeric scores are used, ensuring evaluation remains objective and narrative-focused.

Features

  • 11-seat blind-parallel panel: Each panelist reviews the text independently, without awareness of others' evaluations.
  • Qualitative bands: Assessments use bands (e.g., Superior, Strong, Adequate, Weak) instead of numeric scores.
  • Verbatim evidence: Panelists must cite exact passages from the text to justify their evaluations.
  • Chinese language support: Optimized for Chinese memoir and narrative writing.
  • Claude Code plugin: Integrates directly with Claude Code for seamless workflow.
  • Codex skill: Also available as a Codex skill for broader compatibility.

How It Works

  1. Submit text: Provide a Chinese memoir or narrative text to the panel.
  2. Blind review: Each of the 11 panelists independently evaluates the text.
  3. Qualitative bands: Panelists assign a band (e.g., Superior, Strong, Adequate, Weak) based on literary merit, emotional resonance, structure, and authenticity.
  4. Evidence-based feedback: Each panelist includes verbatim quotes from the text to support their band assignment.
  5. Panel summary: The skill aggregates the results, providing a comprehensive review with insights from all 11 participants.

Use Cases

  • Author feedback: Receive structured, evidence-based critique on a memoir or narrative draft.
  • Literary workshop: Facilitate a group review session for Chinese creative writing.
  • Editing support: Identify strengths and weaknesses in narrative structure, voice, and emotional impact.
  • Educational tool: Teach principles of literary analysis through automated, transparent review.

Requirements

  • Access to a supported AI agent platform (Claude Code or Codex).
  • The skill file (SKILL.md) properly installed.
  • Input text must be in Chinese (simplified or traditional) for optimal performance.

Installation

Refer to the platform-specific documentation for installing skills or plugins in Claude Code or Codex. Ensure SKILL.md is placed in the appropriate configuration directory.

Example

# Sample usage prompt to the AI agent:
Review the following Chinese memoir excerpt using the lit-panel skill:

“我記得那天下著細雨,老街的石板路滑溜溜的。母親撐著一把黑傘,走在前面,背影瘦瘦的,像要融進雨裡。”

The skill will simulate 11 panelists, each providing a band and verbatim quotes.

License

See the original repository for license details.

Related Skills