# Enoki (enokilabs.ai)

Attacks an AI agent's endpoint to find security and safety flaws, then re-runs the attacks on each release to verify fixes.

- Atlas card: https://inite.ai/en/atlas/enokilabs.ai
- Site: https://enokilabs.ai
- For: Teams shipping AI agents who need to test them for security and safety before release
- Categories: security
- Use 0/100 · alive 100/100

## Used for

- Find prompt injection and data leaks in a deployed agent
- Re-run attacks against each fix on every release
- Trigger attack runs from a CI pipeline before shipping

## Price

- freemium (read 2026-10-21 on https://enokilabs.ai/pricing)

## For developers

- MCP server: ai.enokilabs/enoki (namespace; auth on 2026-10-21)
- Read 2026-10-21

## MCP server, measured daily

Server: https://mcp.enokilabs.ai/mcp, checked every day since 2026-10-21 (3 checks).

- Answered: measuring, 3 daily checks so far
- Median first answer: 298 ms
- Sign-in: OAuth (the client signs the user in)
- Method: https://inite.ai/en/atlas/mcp/report

## Your data

- Certifications stated: iso27001, gdpr
- Read 2026-10-21

---

Source: INITE Atlas, https://inite.ai/en/atlas/enokilabs.ai. JSON: https://inite.ai/api/atlas/enokilabs.ai. MCP: https://inite.ai/api/mcp (atlas_lookup).