Adversarial Specs Review — Productivity skill for Claude Code
Adversarially review a Claude-produced overall task decomposition before any per-task implementation planning begins.
How to install Adversarial Specs Review
Installs to ~/.claude/commands/thibautbaissac-react-express-ai-kit-adversarial-specs-review.md
mkdir -p ~/.claude/commands && curl -fsSL https://raw.githubusercontent.com/ThibautBaissac/react-express-ai-kit/HEAD/.claude/commands/adversarial-specs-review.md -o ~/.claude/commands/thibautbaissac-react-express-ai-kit-adversarial-specs-review.md Restart Claude Code, or start a new session, for it to be picked up.
What Adversarial Specs Review does
name: adversarial-specs-review description: Adversarially review a Claude-produced overall task decomposition before any per-task implementation planning begins. argument-hint: "[source-brief-file] [specs-file]" arguments: [brief, specs]
You are an adversarial planning reviewer. Your job is to review the generated overall specs, usually `tasks/specs.md`, before the user starts the per-task `/plan N` pipeline.
Default inputs:
- Source brief:
$brief, ordocs/*.mdif one obvious candi
Alternatives in Productivity
- Reviews Plans — Review all active architecture plans relevant to the current task 15.6k ★
- New Task — Analyze task complexity and create actionable implementation plan 1.5k ★
- Trio — Force a trio review of the task that follows — pm, designer, and engineer weigh in in parallel before any acti 324 ★
Full documentation available on GitHub
View Source RepositoryRelated Skills
Adversarial Task Plan Review
Adversarially review a generated per-task plan file before implementation begins.
Describe Change
You MUST run this as the FIRST step of ANY implementation task. Determines required planning depth before you
Adversarial Execution
Adversarial Execution — an execution-time review gate: two independent strong models must both prove your work
Axiom Planning
TDD-validated implementation planning with a four-reviewer plan-review quality gate - atomic-task plans with e
MCP Tasks Review Task Implementation
Review the implementation of a task before completion
Armature Agent Review
Use after you have used a CLI, MCP server, API, SDK, hosted service, or web app during a task, to submit a sho
Related Agents
Adversarial Judge.Agent
Use to adversarially evaluate a proposal or implementation against its repository profile and specs, returning
Specs
Write a feature specification in docs/specs/ before any implementation begins. Invoke at the start of every ne
Clarifier
Identifies ambiguities and documents assumptions before planning begins. Use this before the planner on any ta