Core concepts

Term	Meaning
Source	One connection Amnify reads: a GitHub organization, an AWS, Azure or GCP connection, or an application you authorize for penetration testing
Scope	One thing inside a source that is scanned on its own: a repository, account, subscription, project or penetration test target
Scan	One execution of one kind of check against one scope
Finding	A confirmed issue. An agent investigates the scan results and writes one finding that states the real impact. It survives re-scans and resolves when the issue is gone
Cluster	Findings that share one root cause, linked together. One remediation covers the whole cluster
Context	A decision your organization has made, captured as a question and your answer, then reused. It decides whether a finding applies to you and what gets fixed first, not how severe the issue is. A finding with an unanswered question is held back until you answer
Knowledge source	A security document you upload for Amnify to read
Remediation	The planned fix for one cluster, with a plan you can edit and approve
Outcome	How a fix lands: a pull request, or an assignment with a runbook

What a finding carries

Impact. What it exposes and why it matters. Not a fix hint.
Severity. Critical through Info, rated against established security frameworks and scoring, not against your preferences.
Location. Repository, file and line, or the cloud resource.
Evidence. What was actually observed.
Status. Open, Resolved or Suppressed, and a closed finding keeps the reason it closed.

Findings come in five types: exposed secrets, code vulnerabilities, infrastructure as code, cloud, and penetration test.