ai-ethics-advisor — Claude Code subagent: setup and tasks
This subagent reviews AI systems for bias, fairness problems, and regulatory gaps. Use it for ethics audits and responsible-AI documentation.
npx claude-code-templates@latest --agent ai-specialists/ai-ethics-advisorWhat ai-ethics-advisor does
ai-ethics-advisor gathers details about the system, deployment jurisdictions, affected groups, decision authority, and existing records. It then examines data and model behavior for demographic or other bias, maps possible harms, and helps produce a model card, algorithmic impact assessment, or ethics review.
The subagent runs on sonnet and can use Read, Write, Edit, Bash, Glob, Grep, WebSearch, and WebFetch. When a report cites a current regulatory date or framework detail, it is instructed to verify the source through search and page retrieval.
For hiring, healthcare, credit, and law enforcement systems, human oversight belongs in the review because these are high-stakes uses. The assessment also covers protected groups, explainability, appeals, and possible harm where those concerns apply.
What it helps with
- audit a model for demographic bias
- identify gaps in AI regulatory compliance
- draft a model card for an AI system
- prepare an algorithmic impact assessment
- review risks affecting protected groups
Example prompts for the agent
- Review our resume screening model for demographic bias before deployment.
- Create an algorithmic impact assessment for our patient triage AI.
- Audit the credit-scoring agent for fairness, human oversight, and tool-access risks.
How to install ai-ethics-advisor
1claude-code-templates
npx claude-code-templates@latest --agent ai-specialists/ai-ethics-advisor2Manually: put the files in place
.claude/agents/ai-ethics-advisor.mdFor all projects use ~/.claude/agents/. Claude delegates by the description field or an @-mention
Manually: put the files in place
.codex/agents/ai-ethics-advisor.tomlCodex defines subagents in TOML: name, description, developer_instructions. Move the Markdown text into developer_instructions by hand. For all projects use ~/.codex/agents/
Manually: put the files in place
.cursor/agents/ai-ethics-advisor.mdCursor also reads .claude/agents/
Step by step, with commands for each platform: Claude Code subagents: how to create one, call it and keep context clean
Questions about ai-ethics-advisor
How do I install ai-ethics-advisor in Claude Code?
Run npx claude-code-templates@latest --agent ai-specialists/ai-ethics-advisor, then add .claude/agents/ai-ethics-advisor.md.
What does ai-ethics-advisor do?
It reviews AI systems for bias, fairness violations, and regulatory gaps. The subagent can also prepare a model card, algorithmic impact assessment, or ethics review.
Which tools does ai-ethics-advisor use?
It uses Read, Write, Edit, Bash, Glob, Grep, WebSearch, and WebFetch, with sonnet as its model.
