robertoecf/adversarial-review

8 stars · Last commit 2026-07-10

Review triad plugin for Claude Code: prompt optimization, adversarial red-team review, plan validation, and intelligent orchestration.

README preview

# adversarial-review

**v0.7.0** — Cross-host adversarial review for coding workflows. Works in
**Claude Code, Codex, Pi, and Grok** from the same source: detects which agent
host you're running in and routes the heavy critique to the OTHER agent.

Critique discipline (v0.7): break confidence rather than validate; expensive
attack surfaces first; material-only findings; steerable focus; terse
ship/no-ship summary. Lessons adapted from OpenAI Codex Companion — see
`references/codex-lessons.md`.

## The principle

> **The partner reviews, never the host.**

Two models examining the same artifact from different angles catch more
issues than either alone. Each has different training biases, blind spots,
and reasoning patterns. Disagreements between the two surface the
highest-value findings - the ones a single reviewer would miss.

View full repository on GitHub →