Skip to main content
Code Audit gives you a single, ranked view of everything worth improving in your project. It runs a suite of open-source scanners against your repository and combines their results with AI-generated improvement ideas, organized by category. From here you can review each item and convert it straight into a task on your board. Screenshot 2026 06 22 At 8 07 31 PM

​
On this page

​
Scanner coverage

A banner at the top summarizes the security scan: how many scanners produced evidence, and the total number of findings broken down by severity (such as high and medium). Expand it to see each scanner individually — with its name, how many findings it produced, and how long it took to run. The scanners are established open-source tools, each strong at a different kind of check:
ScannerWhat it’s best at
banditSecurity issues in Python source code.
gitleaksSecrets and credentials accidentally committed to the repo.
osv-scannerKnown vulnerabilities in your dependencies, matched against the OSV database.
semgrepPattern-based static analysis across many languages.
trivyVulnerabilities across dependencies, containers, and the filesystem.
Running several scanners and consolidating their output means a single problem reported by more than one tool is grouped together — you act on it once instead of chasing duplicate alerts.

​
Ideas and findings

Beyond raw scanner output, Code Audit presents AI-generated improvement ideas for your project, shown as a ranked list with a count of how many ideas exist. Each item is a card showing:
  • Category — what kind of improvement it is (for example, Security).
  • Severity — the priority of the item (for example, high).
  • Status — whether you’ve already acted on it; an item turned into a task is marked converted.
  • Title and description — what was found and where, with a short explanation. Use the open-external icon on a card to view the underlying detail.

​
Filtering by category

Code Audit covers more than security. Use the category tabs to focus the list:
  • All — everything in one list.
  • Code — code-quality and implementation improvements.
  • UI/UX — user interface and experience suggestions.
  • Docs — documentation gaps.
  • Security — vulnerabilities and hardening, backed by the scanners above.
  • Performance — efficiency and speed improvements.

​
Working with findings

The toolbar and per-card controls let you manage the list:
  • Add More — generate additional ideas for the project.
  • Select — use the checkboxes to act on items in bulk.
  • Hide — dismiss items you don’t want to see.
  • Filter and sort — reorder or narrow the list.
  • Refresh — re-run to pick up the latest state.
  • Delete — remove items from the list.
To act on an item, convert it into a tracked task — it moves onto your kanban board and is picked up by the build pipeline. The full flow from a finding to a reviewed pull request is covered in Patch a security vulnerability end-to-end.