Results
What does an AppTruth scan produce?
An AppTruth scan produces evidence-backed behavior findings with confidence, impact, likely cause, affected areas, and fix prompts.
Reviewed by Uriel from AppTruth · Updated · 3 min read
Direct answer to “What does an AppTruth scan produce?”
An AppTruth scan produces a structured behavior report. Each finding can include a plain-language behavior, confidence score, impact level, likely cause, affected areas, file-based evidence, and an implementation-ready remediation prompt.
Key points: What does an AppTruth scan produce?
- Findings describe what the code supports, not just whether a rule passed.
- Evidence points to specific files and observations.
- Pro results include detailed scan coverage and recorded limitations.
The anatomy of a behavior finding
A completed scan returns normalized behaviors rather than a raw model transcript. A behavior explains what the application appears to do. The supporting fields help a user judge how important and trustworthy that conclusion is.
- Behavior: a plain-language statement about the implemented flow
- Confidence: how strongly the selected source supports the conclusion
- Impact: a low, medium, or high indication of potential importance
- Why it happens: the likely implementation cause
- What it can affect: users, data, access, billing, or other product areas
- Evidence: allowed file paths and concise source observations
- Remediation prompt: technical guidance ready for an engineer or coding agent
Expected or not expected?
A surprising finding does not automatically mean the application is wrong. AppTruth lets the user review behaviors and classify whether they match the intended product. If the behavior is correct, the expectation can change. If the behavior is wrong, the evidence and remediation prompt create a focused starting point for a fix.
Coverage and limitations
Eligible Pro scans add coverage details such as reviewed files, detected entry points, omissions, unsupported languages, parse failures, unresolved imports, and static-analysis boundaries. Coverage is deliberately described as broad, partial, or limited rather than exhaustive.
What each AppTruth result field helps a reviewer decide
| Result field | Question answered | Reviewer action |
|---|---|---|
| Behavior | What does the code appear to do? | Compare with intent |
| Confidence | How strongly does source support it? | Review uncertainty |
| Impact | What could be affected? | Prioritize follow-up |
| Evidence | Where is the conclusion grounded? | Inspect the files |
| Remediation | How might it be changed? | Review before implementation |
The BECAUSE Finding Standard
A quality test for behavior findings: a claim should be clear, bounded, reviewable, and actionable.
- Behavior: State the implemented behavior in plain language.
- Evidence: Point to allowed files and concise observations.
- Confidence: Express how strongly the selected source supports the conclusion.
- Affected areas: Name users, data, access, billing, or operations that may be involved.
- Uncertainty: Record coverage and runtime limitations.
- Suggested next step: Offer a reviewable fix or verification action.
Trusted sources that inform this guide
These independent sources support the surrounding verification practices. AppTruth-specific product statements are product guidance and should be confirmed against current account or contractual documentation when formal assurance is required.
- Secure Software Development Framework (SSDF) — NIST. Defines outcome-based secure software development practices, including verification and release preparation.
- Testing for Reliability — Google SRE. Explains why multiple testing methods are needed to build confidence in production systems.
Related questions about “What does an AppTruth scan produce?”
Does a high-impact finding mean the app is definitely wrong?
No. Impact describes potential importance. The owner must still decide whether the behavior is intended and whether runtime validation is needed.
What makes an AppTruth finding reviewable?
The behavior statement is paired with confidence, affected areas, file-based evidence, likely cause, and coverage context.