Skip to content
CONSTABLE
ProblemDefinitionsMechanismFor teamsScopeLexiconImplementation
SDSResearch, diagnosis, and architectural work.ZTGThe open specification defining structural governance requirements.ConstableThe commercial reference implementation and control kernel.

SDS maintains the open ZTG specification and builds Constable as a reference implementation. ZTG can be implemented without Constable.

SHADOW DYNAMIC SYSTEMS

Constable is a product of Shadow Dynamic Systems LLC. Built for institutions that treat authority as infrastructure.

Contact

Jason Crittenden
Founder & Research Lead

jason@shadowdynamicsystems.com

Start here

Related

Zero Trust GovernanceZTG source on GitHubShadow Dynamic Systems
General Constable marketing material. No third-party certification, endorsement, regulatory approval, customer result, or insurance validation is claimed. Constable is a control kernel; it does not by itself make AI risk-free, guarantee compliance, or govern surfaces where it is not integrated.
© 2026 Shadow Dynamic Systems LLCAuthority defines admission.
← Implementation

Irreversibility of Harm

Implementation commentary for ZTG-5 — Irreversibility of Harm ↗ on Zero Trust Governance (v0.7). Constable implements concepts derived from ZTG; this page is commentary, not a conformance claim. Written against snapshot 0.8-draft-2026-09-21 · spec/14-irreversibility-of-harm-ztg-5.md ↗.

Implementation commentary · seed

Constable implements ZTG-5 ↗ on Zero Trust Governance as harm-class and ceiling declarations on its surface registry, a composite-analysis subsystem that bands assessment deterministically, and architecturally distinct gates for high-consequence actions.

Surface and sub-surface registration. Constable registers harm-class defaults and compositional multipliers on surfaces and sub-surfaces in the same ZTG-3 ↗ on Zero Trust Governance surface registry, carried in the ZTG-0e ↗ on Zero Trust Governance governance bundle. Default classifications ship as starting points, not normative classifications: they are conservative defaults a deploying institution is expected to review and tighten for its context, not framework assertions about what a given surface's harm class is.

Liability ceiling computation. Ceilings are computed from deployment-level policy and attached to each authorization request, then written into the ZTG-4 ↗ on Zero Trust Governance evidence record. Constable originates no ceiling itself; the values come from the ratifying principal's policy. Harm-class assignment and ceiling assignment are recorded as HARM_CLASS_ASSIGNED and LIABILITY_CEILING_ASSIGNED events under ZTG-0a ↗ on Zero Trust Governance, so both are observable and replayable as of decision-time.

Composite-analysis subsystem. A composite-analysis subsystem computes assessment as the recursive product of component assessments and surface multipliers, in discrete bands, so the result replays identically through the ZTG-0b ↗ on Zero Trust Governance harness. No continuous-valued computation enters the governance-determining path.

Irreversible-class gates. Constable routes Irreversible-class and over-threshold actions to architecturally distinct high-end gates — separate gate structures, not the ordinary gate with extra checks. HumanSeal presents high-consequence authorizations to operators with the harm class, ceiling, and composite assessment surfaced, so a human engaging an Irreversible-class action sees its gravity rather than an undifferentiated approval prompt.

Critical-class evidence binding. Irreversible-class and over-threshold actions are written as critical-class records on the Monotonic Logger, with the highest durability and integrity the substrate supports, bound write-ahead of the effect per ZTG-4 ↗ on Zero Trust Governance. Mitigable-class actions record their residual harm.

Operator support and conformance evidence. Constable provides operator support for calibrating multipliers and ceilings against experience, as a governed (ZTG-0e ↗ on Zero Trust Governance) activity. Conformance tests cover harm-class assignment, monotonic sub-surface tightening, banded deterministic assessment and its replay, distinct-gate promotion on either sufficient condition, ceiling presence in evidence, and reversal_strategy non-relaxation. The protocol is documented in the conformance verification specification referenced in §22.

PreviousInput Sanitization BoundaryNextMechanistic Boundary