Future boundary
Which unresolved paths are still admissible, which are excluded by evidence, and which remain explicit trust assumptions?
RELIANCE BEFORE CLOSURE · FROZEN INSTRUMENT v0.4.0
Reliance Inspector
A workflow can still be open after one decision has become stable enough to rely on.
Reliance Inspector shows the evidence, assumptions, semantic identities, stability margin, and action bound behind that narrower decision. It lets you inspect why a relying result is SUPPORTED, why it must be REQUALIFY, or why matching numbers still become NOT_APPLICABLE.
Verdict stability
Across that qualified future boundary, how much can the claim move before its verdict changes?
Action effect
How much can the proposed action itself worsen the same claim, including qualified post-action expansion?
Relying identity
Do the certificate, action bound, and relying party agree on the exact claim, unit, boundary, profile, and trust identity?
Provider-blind seam
qualified boundary + stable claim + qualified action + exact identity
rely()SUPPORTEDREQUALIFYNOT_APPLICABLE
SUPPORTED means semantic applicability, not permission. It does not infer authorization, reserve capacity, or guarantee a concurrent commit.
PUBLIC CANONICAL MODE
Recorded commissioning evidence, negative controls, and counterfactual cases are inspectable in-browser from the frozen v0.4.0 case data.
Local verification boundary
Evidence-bundle verification and semantic-object evaluation require the archived loopback server and frozen Python consumer. Windows portable · Archived v0.4.0 · Source
Selected canonical case
Relying result
semantic applicability
Decision trace
Why this result follows
Each stage answers a different question. No later stage repairs an unqualified earlier one.
Recorded timing
Claim-specific reliance before broader closure
The interval is one commissioning measurement, not a performance average.
reliance point
measured interval
broader closure
Boundary status
Stability margin M
Action bound Cₐ
Reason codefrozen consumer
Interpretation
Observed facts
Evidence-backed exclusions
Paths removed by recorded evidence rather than assumption.
Explicit trust assumptions
Accepted conditions that remain part of the claim boundary and are not promoted into evidence.
Inspect immutable canonical case JSON
LOCAL EVIDENCE VERIFICATION
Reconstruct the decision from an evidence bundle
The full local release verifies a compatible recorded evidence ZIP, rebuilds the claim-specific boundary and derivations, and emits a deterministic verification record. The public website deliberately does not receive or process that bundle.
Recheck the evidence manifest and bundle identity.
Separate evidenced exclusions from explicit trust assumptions.
Reconstruct the margin, qualified action bound, and relying result.
Emit exact evidence, verifier, and semantic-core identities.
LOCAL SEMANTIC EVALUATION
Ask whether already-verified semantic objects apply
The local evaluator consumes a stability certificate, qualified action bound, and relying context. It requires exact semantic and trust identity before applying the frozen threshold-slack rule.
PreconditionThis mode does not verify signatures, issuers, authenticity, or transport integrity. The supplied semantic objects must already have been verified upstream.
Claim, profile, unit, boundary, and trust identity must match.
For stable SATISFIED, support requires Cₐ ≤ M. The VIOLATED case uses the strict bound.
Return SUPPORTED, REQUALIFY, or NOT_APPLICABLE without inferring authorization or reservation.