Glossary
Every term used on this site, in one place.
- Application Brain
- The facts AppScan keeps about your application between scans, each with a confidence and a provenance.
- Change set
- Pushes to one branch, coalesced into a single unit of work, classified by what they touched.
- Coverage
- What a scan actually managed to look at: steps completed, failed and skipped.
- Credit
- The unit AI work and manual scans are priced in. Plans include a number per usage window.
- Confidence
- How certain a detector is, between zero and one. It affects ranking, never whether you are told.
- Correlation
- Joining a runtime observation to the source that caused it, so one problem is one finding.
- Domain
- What kind of problem a finding is: security, privacy, performance, cost abuse or reliability.
- Entitlement
- A limit or capability your plan grants, enforced by the API rather than described in marketing.
- Evidence
- What a rule collected to justify a finding, already masked.
- Fingerprint
- A finding’s semantic identity, so it survives a refactor and does not duplicate.
- Finding
- One problem, with five answers, a rule id, a location and evidence.
- Invariant
- Something true about your code that should stay true, derived from the code itself and re-checked on every commit.
- NEW / UNCHANGED
- Whether a finding on a pull request was introduced by that change or inherited from its base branch.
- Partial coverage
- A scan that could not finish. Reported as such, never as clean.
- Probe
- The one active check: a bounded, read-only replay that proves whether a data API answers anyone.
- Rule id
- The stable identifier of the check that raised a finding, such as source.injection.sql-built-from-input.
- Severity
- Impact if the finding is real: critical, high, medium, low or info.
- Site
- One application: its address, policy, history, repositories and findings.
- Taint analysis
- Following request input through the code to the sink it reaches, rather than matching on dangerous-looking names.
- Usage window
- The period a plan’s included credits belong to.
Questions
What does “deterministic” mean here?
The same input produces the same finding every time, with no model in the loop. It is what makes a result reproducible and arguable.
Something here wrong, missing or out of date? Tell us and we will fix the page.