Now liveThe Skillselion MCP - thousands of ranked skills, loaded into your agent mid-task. No install.Get it →
Audn.ai avatar

Audn Security Tools

  • 6 repo stars
  • Updated April 13, 2026
  • audn-ai/skills

Audn Security Tools is a Claude Code plugin marketplace that orchestrates Audn.ai adversarial campaigns so developers can red-team text and voice AI agents for jailbreaks, injection, and compliance gaps.

About

Audn Security Tools is a Claude Code plugin marketplace for adversarial security testing of AI agents. Skills orchestrate the Audn.ai API to register agents, run attack campaigns, and deliver graded vulnerability reports with remediation guidance. Developers reach for it when they need to red-team chat or voice agent endpoints from the IDE before shipping. Install with npx skills add or /plugin marketplace add audn-ai/skills, authenticate via /audn-login, then run /audn-red against an API URL.

  • Automated red-team campaigns for text and voice AI agent endpoints
  • Attack library covers jailbreaks, prompt injection, and data extraction
  • Bundled OAuth MCP server with 34 Audn API tools for Claude Code
  • Compliance reports for ISO 42001, SOC2, executive, and vulnerability views

Audn Security Tools by the numbers

  • Data as of Jul 10, 2026 (Skillselion catalog sync)
/plugin marketplace add audn-ai/skills

Add your badge

Show developers this marketplace is listed on Skillselion. Paste this into your README.

Listed on Skillselion
repo stars6
Last updatedApril 13, 2026
Repositoryaudn-ai/skills

How do I adversarially test my AI agent API for jailbreaks, prompt injection, and data leakage before release?

Red-team voice and text AI agents from Claude Code with automated jailbreak, prompt-injection, and compliance campaigns via Audn.ai.

Who is it for?

Developers building LLM or voice agents in Claude Code, Cursor, or Codex who need automated red-team audits and compliance reports from the IDE.

Skip if: Teams testing only traditional web apps without AI endpoints or those unwilling to use an Audn.ai account and API access.

What you get

Delivers security grades, vulnerability tables, attack category breakdowns, remediation recommendations, and optional ISO 42001 or SOC2 reports.

Recommended Marketplaces

FAQ

How do I authenticate in Claude Code?

Run /audn-login, restart Claude Code, authenticate the audn-redteam MCP server via browser OAuth, and tokens are stored in your system keychain.

What does /audn-red do?

It registers your agent, selects attack scenarios, executes a campaign, monitors progress, and returns a full vulnerability report with an A-F security grade.

Which attack categories are covered?

Campaigns test prompt injection, jailbreaks, data extraction, social engineering, emotional manipulation, and role confusion.

Securityauditappsec

This week in AI coding

Five minutes, every Monday - the tools, releases and tactics for developers.

unsubscribe anytime.