coreybrown

Ethics Trust Safety — Development agent for Claude Code

Development community

The adversarial "how does this harm or get weaponized" pass — dark patterns, exploitation of user attention, abuse vectors, bias, and moderation burden.

How to install Ethics Trust Safety

Installs to ~/.claude/agents/coreybrown-proposal-engine-ethics-trust-safety.md

Terminal

mkdir -p ~/.claude/agents && curl -fsSL https://raw.githubusercontent.com/coreybrown/proposal-engine/HEAD/.claude/agents/ethics-trust-safety.md -o ~/.claude/agents/coreybrown-proposal-engine-ethics-trust-safety.md

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

What Ethics Trust Safety does

name: ethics-trust-safety description: The adversarial "how does this harm or get weaponized" pass — dark patterns, exploitation of user attention, abuse vectors, bias, and moderation burden. Findings-only risk agent in Stage 4 of the /propose pipeline. tools: Read, Write, Glob, WebSearch model: sonnet

You are the ethics and trust/safety auditor. Legal asks "can we?"; you ask "should we?" — plenty of legal features are dark patterns. You have the most explicit license to be negative of

Alternatives in Development

Auth Tester — Authentication and session management testing agent 812 ★
Executioner — Optional Sanhedrin fallback verifier 632 ★
Cold Verifier — Phase 11 cold verification agent that independently re-verifies CRITICAL and HIGH findings with zero prior con 125 ★