ClawKit Logo
ClawKitReliability Toolkit
Back to Registry
Official Verified developer tools Safety 4/5

adversarial-coach

Adversarial implementation review based on Block's g3 dialectical autocoding research. Use when validating implementation completeness against requirements with fresh objectivity.

Why use this skill?

Use the adversarial-coach to validate code against requirements. Automate rigorous implementation reviews, catch security gaps, and ensure total compliance with your project specs.

skill-install — Terminal

Install via CLI (Recommended)

clawhub install openclaw/skills/skills/killerapp/adversarial-coach
Or

What This Skill Does

The adversarial-coach skill implements a dialectical autocoding feedback loop based on Block’s g3 research. It functions as an objective, adversarial reviewer that audits your code against specified requirements. Instead of merely suggesting improvements, it acts as a gatekeeper, validating your implementation's completeness, security, and edge-case resilience. By forcing a separation between the implementing 'player' agent and the reviewing 'coach', it mitigates cognitive bias where an AI might otherwise rationalize its own shortcuts or ignore missing requirements. The process terminates only upon the issuance of the IMPLEMENTATION_APPROVED magic signal.

Installation

To integrate this skill into your environment, use the OpenClaw CLI:

clawhub install openclaw/skills/skills/killerapp/adversarial-coach

Use Cases

Use this skill during high-stakes development cycles where requirements compliance is paramount. It is ideal for:

  • Security-hardened backends where authentication and encryption protocols must be strictly verified.
  • Complex feature development with multiple interconnected requirements.
  • Final QA passes before merging code into a production environment.
  • Preventing 'lazy' coding patterns where the agent assumes requirements are met without providing tangible test evidence.

Example Prompts

  1. "/coach requirements.md"
  2. "I've completed the implementation of the OAuth flow and the user profile schema. Run /coach to verify my work against the project specs."
  3. "/coach"

Tips & Limitations

The adversarial-coach is designed to be rigorous. It will not approve code that is 'almost' there; it requires 95%+ completion and passing tests. When utilizing this skill, ensure your requirements.md or SPEC.md files are clearly structured. The coach ignores style preferences; do not attempt to use it for linting or formatting concerns. Its primary strength lies in identifying security gaps like missing token revocation, lack of input sanitization, or unhandled 401/403 scenarios. If the coach returns a list of immediate actions, address them systematically before running the command again to reach the approval signal.

Metadata

Author@killerapp
Stars1776
Views2
Updated2026-03-02
View Author Profile
AI Skill Finder

Not sure this is the right skill?

Describe what you want to build — we'll match you to the best skill from 16,000+ options.

Find the right skill
Add to Configuration

Paste this into your clawhub.json to enable this plugin.

{
  "plugins": {
    "official-killerapp-adversarial-coach": {
      "enabled": true,
      "auto_update": true
    }
  }
}

Tags(AI)

#code-review#qa#adversarial-testing#automation#security
Safety Score: 4/5

Flags: file-read, code-execution

Related Skills

agentskills-io

Create, validate, and publish Agent Skills following the official open standard from agentskills.io. Use when (1) creating new skills for AI agents, (2) validating skill structure and metadata, (3) understanding the Agent Skills specification, (4) converting existing documentation into portable skills, or (5) ensuring cross-platform compatibility with Claude Code, Cursor, GitHub Copilot, and other tools.

killerapp 1776

aws-agentcore-langgraph

Deploy production LangGraph agents on AWS Bedrock AgentCore. Use for (1) multi-agent systems with orchestrator and specialist agent patterns, (2) building stateful agents with persistent cross-session memory, (3) connecting external tools via AgentCore Gateway (MCP, Lambda, APIs), (4) managing shared context across distributed agents, or (5) deploying complex agent ecosystems via CLI with production observability and scaling.

killerapp 1776

baml-codegen

Use when generating BAML code for type-safe LLM extraction, classification, RAG, or agent workflows - creates complete .baml files with types, functions, clients, tests, and framework integrations from natural language requirements. Queries official BoundaryML repositories via MCP for real-time patterns. Supports multimodal inputs (images, audio), Python/TypeScript/Ruby/Go, 10+ frameworks, 50-70% token optimization, 95%+ compilation success.

killerapp 1776

chain-of-density

Iteratively densify text summaries using Chain-of-Density technique. Use when compressing verbose documentation, condensing requirements, or creating executive summaries while preserving information density.

killerapp 1776

copywriter

Write compelling UX copy, marketing content, and product messaging. Use when writing button labels, error messages, landing pages, emails, CTAs, empty states, tooltips, or any user-facing text.

killerapp 1776