17. Annex E: External Reference Boundary and Evidence Alignment Rules

External standards, frameworks, methods, and tools may be referenced in future DACS guidance. Unless explicitly incorporated into a normative DACS control, such references are informative by default.

• No hidden conformance. A claim of alignment with another standard does not automatically satisfy a DACS control.
• No automatic substitution. A tool-generated artifact satisfies a DACS control only when it contains the evidence required by that control.
• Implementation neutrality. DACS does not require a specific tool, platform, schema, graph, model, workflow engine, or repository.
• Evidence boundary. External frameworks may provide evidence inputs; DACS defines the criteria and result vocabulary used to evaluate those inputs for the stated decision scope.

Guardrail boundary. AI guardrail systems, policy engines, retrieval filters, model monitors, security controls, runtime validation mechanisms, and similar mechanisms may support DACS-relevant checks or produce DACS evidence. Their label or existence does not satisfy a DACS control. They satisfy a control only when they record what it requires: evidence, scope binding, Gate Record, resulting readiness state, and exception or violation consequence.

DACS is designed to operate alongside adjacent standards, frameworks, and tools. Adjacent references may supply evidence for data quality, metadata, controlled vocabulary, provenance, graph constraints, lineage, AI risk, AI management systems, audit controls, security/privacy controls, process maturity, or governance operating models. DACS defines the criteria used to evaluate that evidence against the applicable decision-scope controls. The assessor or governed control process applies those criteria and records the control result.

DACS focuses on decision-readiness boundaries that can remain between adjacent disciplines when their artifacts are used for material decision paths:

• business meaning is not physically anchored to source systems, attributes, identifiers, semantic grain, and resolution rules;
• relationship meaning is not physically anchored to relationship evidence, orientation, vectors, permitted paths, and constraints;
• discovered, inferred, generated, imported, or retrieved evidence is treated as ready for stated use before Gate controls pass;
• AI outputs are evaluated without a governed context contract, permitted relationship paths, drift rules, or enforceable consequences;
• technical lineage exists but the decision path cannot be reconstructed from identity, meaning, relationship, context, Gate, and evidence records;
• governance documentation exists but cannot be observed, tested, evidenced, or translated into a Gate readiness state;
• Capability Level is confused with binary control conformance for a stated decision scope.

These evidence-alignment rules are informative. Future DACS reference-alignment guidance may map DACS controls to external standards and frameworks, but no external artifact satisfies a DACS control unless the required DACS evidence exists.