plan-grader

Community

Grade implementation plans 0-10 safely, no workspace changes.

Authorkaiohenricunha
Version1.0.0
Installs0

System Documentation

What problem does it solve?

Manually evaluating agent-generated implementation plans for completeness, safety, and alignment with project rules risks accidental workspace changes and inconsistent scoring. This Skill eliminates that risk with a fully non-mutating, evidence-based grading workflow.

Core Features & Use Cases

  • Weighted Rubric Scoring: Grades plans on 7 evidence-based dimensions (scope clarity, intent alignment, repo grounding, execution order, harness compatibility, validation, and risk handling) for an objective 0-10 score.
  • Multi-Agent Support: Accepts plan files, inline quoted text, or auto-discovers latest plans from Claude Code, Codex, GitHub Copilot, and Gemini CLI.
  • Blind Scoring Mode: Strips agent attribution to eliminate bias when comparing plans from different tools.
  • Use Case: Engineering leads can quickly validate implementation plans from their team's agentic CLIs to catch missing validation, harness violations, or risky changes before coding begins.

Quick Start

Use the plan-grader skill to score your latest Codex implementation plan and receive a list of blocking gaps to address before running the plan.

Dependency Matrix

Required Modules

None required

Components

references

💻 Claude Code Installation

Recommended: Let Claude install automatically. Simply copy and paste the text below to Claude Code.

Please help me install this Skill:
Name: plan-grader
Download link: https://github.com/kaiohenricunha/dotbabel/archive/main.zip#plan-grader

Please download this .zip file, extract it, and install it in the .claude/skills/ directory.
View Source Repository

Agent Skills Search Helper

Install a tiny helper to your Agent, search and equip skill from 536,000+ vetted skills library on demand.