← All use cases

Run your first security scan

Scan your code for weaknesses, weigh the evidence, and choose which issues to act on first.

ZeroTwo's Security plugin scans a repository, lets you inspect the evidence and how much was covered, and helps rank findings for follow-up or repair.

Intermediate
Long-running

Skills & Plugins

  • Security: Security Scan
    Runs a threat model, finding discovery, validation, and attack-path analysis, then writes a report with coverage ($security-scan).

Starter prompt

$security-scan Run a scan on the api/ directory of this repository, which I'm authorized to review. Report what you scanned and which commit, then list the findings ordered by how reachable and exploitable they are, with the evidence behind each. Don't modify any files and don't mark anything as resolved.

Before you start

Enable the Security plugin from Customize, then open a Work or Code task on a repository you are authorized to review. Start with one repo or one service folder rather than a whole monorepo.

Run your first scan

  1. Invoke the skill

    In the composer, type $ and pick security-scan, or paste the starter prompt.

  2. Set the scope

    Name the repository or a path such as api/. A scoped scan stays inside that folder and the files that support it.

  3. Approve subagents

    The scan splits discovery and validation across subagents, so it asks for your permission to use them before it starts.

Follow the scan and check coverage

The scan sets a goal for itself and keeps working until every in-scope file has been covered or explicitly deferred. Watch progress in the task, then confirm the target, the commit and any skipped areas in the final markdown and HTML reports. An empty list with incomplete coverage proves nothing.

Understand a finding

For each finding, ask what an attacker can do, how they reach it and what validation showed. Severity alone isn't enough.

Choose the next action

Record each review, then fix via Fix a finding from your security scan, investigate further, or close with evidence. For depth see Run a deep security scan.

Related use cases