scaffold-review

Analyze agent conversation history to identify gaps and drift in AGENTS/CLAUDE instructions.

3|2|Updated Sep 3, 2025
One-click install
npx skills add https://github.com/ishandhanani/dotfiles --skill scaffold-review
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: scaffold-review
Source: https://github.com/ishandhanani/dotfiles/tree/main/agents/skills/scaffold-review
Command: npx skills add https://github.com/ishandhanani/dotfiles --skill scaffold-review

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Identify and remedy gaps, drift, and misalignment in AGENTS/CLAUDE instructions and skills to improve consistency and performance.

Core Features & Use Cases

  • Analyze conversation histories to surface drift in prompts, tool usage, and skill adherence.
  • Propose concrete scaffold improvements and apply changes with governance-friendly ledger updates.
  • Track outcomes over time and provide a repeatable process for continuous improvement.

Quick Start

Run scaffold-review against the latest agent conversations to generate improvement proposals and apply approved changes.

Frequently Asked Questions about scaffold-review

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

FAQPage Schema
How do I identify and fix scaffold drift in Claude agent instructions?

Scaffold drift is identified by analyzing agent conversation histories to surface misalignments in prompts, tool usage, and skill adherence. The review process proposes concrete scaffold improvements and applies changes with ledger updates.

What is the best way to analyze agent conversation history for prompt misalignment?

Use scaffold review to analyze prior agent runs and surface drift in prompts, tool usage, and skill adherence. This generates a prioritized list of fixes and updated artifacts to realign agent instructions.

How do I track scaffold improvements and apply changes with governance updates?

Track scaffold improvements by applying targeted changes to AGENTS/CLAUDE instructions and recording them in a scaffold ledger, providing a repeatable process for continuous improvement and tracking outcomes over time.

Do I need access to prior runs and a scaffold ledger to review agent instructions?

Yes, reviewing agent instructions requires access to AGENT_HOME, prior runs, and the scaffold ledger to analyze histories and output a prioritized list of fixes and updated artifacts.

Can I use scaffold review to generate a prioritized list of fixes for agent skills?

Yes, scaffold review generates a prioritized list of fixes and updated artifacts by analyzing conversation histories to identify gaps and drift in agent skills, improving consistency and performance.