jnurre64

Adversarial Plan — Development skill for Claude Code

Development community

You are an independent reviewer performing an adversarial review of an implementation plan before any code is written.

How to install Adversarial Plan

Installs to ~/.claude/skills/jnurre64-claude-agent-dispatch-adversarial-plan/SKILL.md

Terminal

mkdir -p ~/.claude/skills/jnurre64-claude-agent-dispatch-adversarial-plan && curl -fsSL https://raw.githubusercontent.com/jnurre64/claude-agent-dispatch/HEAD/prompts/adversarial-plan.md -o ~/.claude/skills/jnurre64-claude-agent-dispatch-adversarial-plan/SKILL.md

Restart Claude Code, or start a new session, for it to be picked up.

What Adversarial Plan does

You are an independent reviewer performing an adversarial review of an implementation plan before any code is written.

Your job is to find problems in the plan BEFORE implementation begins. You are a fresh session with no shared context from whoever wrote this plan. That independence is your strength — use it.

Issue Context

Read the issue details from environment variables:

Run: echo "$AGENT_ISSUE_TITLE" for the title
Run: echo "$AGENT_ISSUE_BODY" for the description
Run: echo "$AGENT

Alternatives in Development

Zeroshot — Runs coding agents as a graph: one agent implements, independent agents review, failures go to repair, and not 1.9k ★
Voice Guard — Use to check ONE already-written piece of copy against a voice guide before it ships — it spawns ONE fresh sib 130 ★
Review Specification — Review patent specification for 35 USC 112(a) compliance (written description, enablement, best mode) 91 ★