Public framework stack and adjacent regimes
InferensLab publishes the intent, boundary, and public role of its framework stack. The stack is visible enough to orient interpretation, but not detailed enough to clone the method.
These frameworks are complementary. They should be read as a sequence of public functions rather than as isolated labels.
Framework companion files
Framework stack
Each framework has its own public role and boundary; the index ties them together.
/.well-known/ai-frameworks.json— Framework-stack index./.well-known/dualweb.json— Dual-surface publication posture./.well-known/ssa-e.json— Evidence and authority discipline.
Action + legitimacy
A2 and QLayer connect framework summaries to publication updates and response authorization.
/.well-known/a2.json— Audit-to-action loop./.well-known/qlayer.json— Response-legitimacy gate./.well-known/ai-governance.json— Top-level router that places the framework stack.
How the frameworks relate
- DualWeb defines the publication posture: one human-facing layer and one machine-first layer, both aligned.
- SSA-E governs evidence and authority discipline without exposing scoring or private protocol logic.
- A2 converts findings into governance signals and controlled publication updates.
- QLayer governs whether a response may be produced at all, and what the valid fallback must be when conditions fail.
- Authority Governance (Layer 3) remains adjacent: it concerns executable permission, not merely interpretive legitimacy.
Public framework summaries
DualWeb
Dual-surface publication posture: human doctrine plus machine-first governance signals, aligned to reduce ambiguity.
SSA-E
Evidence and authority discipline published as intent and boundary, not as reproducible audit machinery.
A2
Audit-to-Action loop: findings become governance signals, routing updates, and publication decisions.
QLayer
Response legitimacy layer. Clarification and legitimate non-response are first-class outputs when the interpretive conditions are not met.
Adjacent regime boundary
Authority Governance (Layer 3) is not the next step of open-web explanation. It is the adjacent regime for delegated execution once an interpretive output becomes action-bearing. InferensLab may describe that boundary publicly, but it does not publish execution permissions, gates, or tooling.
What remains private
- Decision thresholds, scoring schemes, audit choreography, and calibration rules.
- Operational pipelines, scripts, runbooks, and response playbooks.
- Client-specific implementations and internal delivery logic.