Trust States
A reference for evidence-based classification.
01. What Trust States Are
Trust states in SATE-Laravel are descriptive labels that categorize the depth and reliability of the evidence provided by a test.
They are not goals to be achieved or certifications of code quality. A trust state identifies the level of proof the engine was able to extract during analysis. They help developers and architects reason about the strength of their verification signals without implying code correctness or safety.
02. UNTRUSTED
Classification: Insufficient Evidence
The engine classifies a test as UNTRUSTED when the evidence provided is missing, invalid, or unanchored. At this state, the engine cannot establish a provable link between the test's actions and its observations.
This is the default and expected starting state for any test undergoing verification. It represents an absence of conclusive data rather than a definitive failure of logic.
03. STRUCTURAL
Classification: Valid Shape
The test satisfies the foundational requirements of existence and executable shape. The engine can uniquely identify and parse the test structure, confirming that it follows recognized framework patterns.
/ DOES NOT imply correctness.
04. OBSERVABLE
Classification: Evidence Exists
The test successfully targets and measures external or persistent effects in the system state. Physical proof of a side effect exists within the execution context.
Observation alone is fragile. It confirms that a change occurred, but the cause of that change is not yet proven to originate from the test's specific act.
05. CAUSAL
Classification: Attributable Proof
The observed effects are uniquely traceable to the specific Act defined in the test. The Auditor has removed ambiguity regarding the origin of the evidence, establishing a direct causal link.
While confidence is increased, intent is still not verified and logical correctness is not guaranteed.
06. STRICT
Classification: Independent Confirmation
The behavioral behavior is corroborated by multiple independent points of confirmation. This represents the highest depth of behavioral proof currently available to the engine.
/ HIGHER TRUST != SAFETY
/ HIGHER TRUST != ABSENCE OF BUGS
07. What Trust States Do NOT Mean
Trust states categorization is a technical description of evidence quality, not a value judgment.
SATE-Laravel explicitly denies that these states represent:
- Approval
- Certification
- Compliance
- Production Readiness
“Trust states describe evidence, not outcomes.”
These states are intended to help technical users reason about the depth of their evidence. They do not replace developer judgment and they do not absolve any individual of responsibility for the logic under test.