About Sensei (Pro)

Note: Sensei is a DefectDojo Pro-only feature and is currently in BETA.

Sensei is DefectDojo’s AI-powered scan-and-fix capability for source code repositories. Connect a repository (through a GitHub App, GitLab, Bitbucket, or Azure DevOps) and Sensei scans it, imports the results as DefectDojo findings, and then uses a large language model to remediate those findings by opening pull/merge requests, all without leaving DefectDojo.

🔀 Multiple providers: Sensei supports GitHub (github.com and GitHub Enterprise Server), GitLab (gitlab.com and self-managed), Bitbucket (Cloud and Server/Data Center), and Azure DevOps, all with the same scan-and-fix flow. Where this guide says pull request, GitLab uses a merge request; the PR status check is posted as a GitLab/Azure commit status or a Bitbucket build status. Connection differs by provider (see Set up Sensei); everything after onboarding is identical.

• Scan-and-fix in one place: repositories are scanned and remediated from the Sensei page and from your findings, using the same normalized, deduplicated finding data as the rest of DefectDojo.
• Preview-first: Sensei stages fix candidates for review. Nothing is sent to an LLM and no pull request is opened until you approve, so there is no surprise cost or unexpected PR.
• Short-lived credentials: Sensei runs entirely through a GitHub App and uses short-lived installation tokens. There is nothing to paste and nothing to rotate.
• Metered and license-gated: Sensei is a Pro feature with per-instance quotas for fixes and onboarded repositories.

🔎 BETA: Sensei is under active development and is labeled BETA throughout the UI. Behavior and screens may change between releases.

📍 Where to find it: open Sensei from the left-hand navigation.

How DefectDojo-hosted scanning works

DefectDojo-hosted scanning is the recommended way to run Sensei. Scans run inside DefectDojo, and nothing is added to your repository:

1. Connect a GitHub App and install it on the organization (or account) that owns your repositories.
2. Onboard a repository for hosted scanning and choose how findings are reported and (optionally) auto-fixed.
3. Sensei scans the repository (on demand, or automatically when a pull request is opened) and imports the results into an engagement named after the branch.
4. Sensei remediates findings by generating a fix and opening a pull request against the repository’s default branch.

Each onboarded repository is linked to a DefectDojo asset (product), so its findings, engagements, and fixes live alongside the rest of your data.

The three ways a fix gets started

Sensei can remediate a finding in three ways:

• The Fix button on a finding: trigger a one-off fix directly from the findings table or a finding’s detail page. See Fixing findings with Sensei.
• Auto-fix candidates: after each scan, Sensei stages the findings that match your criteria as candidates. You review them and approve the ones to fix (or let Sensei remediate them automatically). See Auto-fix candidates.
• A /fix comment on a pull request: comment /fix on a pull request and Sensei pushes a remediation to that PR.

Requirements

• A DefectDojo Pro license that includes the Sensei feature.
• A connected source-control provider (see Set up Sensei): a GitHub App (github.com or Enterprise Server), a GitLab project/group access token (gitlab.com or self-managed), a Bitbucket connection (Cloud or Server/Data Center — OAuth, API token, or access token), or an Azure DevOps Personal Access Token.
• To configure Sensei (connect apps, onboard repositories): a global Maintainer or Owner role.
• To trigger a fix on a finding: at least Writer access to that finding’s product.

Quotas

Sensei is metered against your license. The Sensei hub shows two usage meters at the top of the page:

• Fixes: the number of remediations applied against your prepaid limit. Approving a candidate or triggering a fix consumes from this quota.
• Onboarded Repositories: the number of repositories onboarded against your repository limit.

When a quota is reached, Sensei blocks further fixes (or onboarding) until it is raised. See Reference for details.