skills.rest
anortham
anorthamCOMMUNITY
Explore all skills by this author →
GitHub

claude-cli

Start fresh Claude Code CLI sessions for independent code reviews.

2|Updated Mar 3, 2026
Search Tags:#cli#claude#code-review#second-opinion#adversarial-review#fresh-session
One-click install
npx skills add https://github.com/anortham/razorback --skill claude-cli-anortham
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: claude-cli
Source: https://github.com/anortham/razorback/tree/main/skills/claude-cli
Command: npx skills add https://github.com/anortham/razorback --skill claude-cli-anortham

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Provides a workflow to obtain a fresh Claude Code CLI session for independent code reviews and adversarial checks. It enables contrastive evaluations by starting from a clean prompt context, reducing cross-talk from prior runs and improving objectivity in assessment.

Core Features & Use Cases

  • Fresh Claude session via claude -p for second opinions, code review, and adversarial review.
  • Read-only review prompts with structured outputs and optional JSON-schema validation for reproducible results.
  • Ephemeral, non-persistent sessions and model routing guidance to align with project policies and cost controls.

Quick Start

Run a fresh Claude review session on your project using claude -p with non-persistent mode and the recommended flags.

Frequently Asked Questions about claude-cli

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

FAQPage Schema
How do I get a fresh Claude session for an independent code review?▼

A second-opinion code review uses a fresh, non-persistent Claude CLI session to independently evaluate your code. This clean context reduces cross-talk from prior runs, providing a contrastive and objective assessment of your work.

How do I run an adversarial code review using the Claude CLI?▼

You can run an adversarial code review by initiating a fresh Claude CLI session in code-adversarial review mode. The workflow uses read-only review prompts to challenge your code from a new perspective, ensuring a rigorous second opinion.

Do I need an Anthropic API key to use Claude CLI for code reviews?▼

Yes, obtaining a fresh Claude CLI session requires Anthropic OAuth or an API key for authentication. The workflow also enforces model routing from the repo-root policy to manage costs and align with project policies.

Can I enforce structured outputs for reproducible code review results?▼

Yes, you can enforce structured outputs for reproducible code reviews using optional JSON schema validation. This manages cost and turn limits while ensuring your read-only review prompts return consistent, structured data.

Why should I use a non-persistent Claude session for a second opinion?▼

A non-persistent Claude session provides a clean prompt context for your second opinion, reducing cross-talk from prior runs. This contrastive evaluation improves objectivity by ensuring the review is not influenced by previous session history.

Related Skills

Explore More Software Engineering Skills→
JuanMS20JuanMS20

github-auth

Configure GitHub authentication for Hermes Agent using git, gh CLI, or SSH credentials.

Community
Intermediate
maslennikov-igmaslennikov-ig
243

parse-error-logs

Parse npm, pnpm, TypeScript, Jest, and build logs into structured JSON reports.

Community
Intermediate
jonnymuirjonnymuir
1

Model Selection

Select optimal LLM models via hierarchical resolution and config.json overrides.

Community
Intermediate
jinwoonghongjinwoonghong

moai-workflow-loop

Integrate LSP diagnostics and AST-grep scanning into an automated code quality feedback loop.

Community
Advanced