Four More Scenarios Make Correction, Eligibility, and Course Echo Inspectable

Release at a glance

The public Scenario library now contains eleven reviewed entries. Four additions widen both the subject matter and the proof model: one formalizes an existing Grid FYI product demonstration as a first-class Scenario, while three show how consequential public or research decisions can remain inspectable when facts, rules, claims, and corrections change.

The release does not turn authored stories into case studies. Each page identifies its mode, evidence ceiling, authority boundary, source revision, fixture stage, and next proof requirement. Each corresponding package retains the canonical claims, gaps, authority model, acceptance contract, fixture outline, authored oracle, diagrams, and derivative records needed to keep future films, presentations, scripts, briefs, and interactives aligned.

The four additions

1. Course Echo: One Transport Disappears connects an existing deterministic TypeScript publication fixture, the Change One Fact lab, and the released film Every System Works. The System Does Not. to one reviewed Scenario contract. Six available transports become five; the original plan loses its margin; two authored branches are compared; and the model stops before resource or command authority. Its product-demonstration level applies to inspectable behavior in this repository’s publication experience, not to a Grid runtime or defense integration.
2. The Finding That Changed After Publication follows a wholly synthetic research package after fifteen generated records are corrected. The exact denominator, table, prose, slide, review state, and successor release change together while the earlier release remains reconstructable. The trace distinguishes computational reproducibility, method review, interpretation, and publication authority.
3. The Warning That Had to Be Corrected Everywhere follows a fictional public warning across web, call-center, translated, partner, archive, and correction surfaces. It keeps policy ownership, message approval, channel publication, receipt, supersession, and observed public outcome as different records instead of treating one successful push as proof that every surface is current.
4. The Award List the Assistant Could Not Approve tests an AI-generated grant recommendation against hard gates, unresolved exceptions, ranked planning states, a funding constraint, and named human authorities. The assistant can propose and explain; it cannot resolve facts, interpret exceptions, obligate funds, issue a determination, or prove that notice was delivered.

A clearer proof ladder

Every Scenario now includes a visible proof-and-limits panel before its capability ledger. The panel answers six questions without asking the reader to infer maturity from polish:

• Is the story public?
• What source-readiness stage has it reached?
• Is its fixture only authored, or does retained execution evidence exist?
• Has an independently identified reviewer evaluated the oracle?
• Is there any deployment evidence?
• What proof must be added next?

The founding seven and the three new synthetic traces remain authored fixture outlines. That is intentional and visible. Course Echo now retains passing execution evidence, an independently reviewed oracle, negative-case coverage, and verification for publication fixture 2026-08-27.2, so its Scenario-owned lifecycle reaches R3. That verification remains limited to the Grid FYI publication trace; it is not Grid-runtime, defense-integration, deployment, execution, or outcome evidence. Public availability and fixture verification remain separate states.

One source, controlled derivatives

The Scenario package is the control point for reuse. Short and long spoken scripts and an accessible brief source are recorded against the same source revision, fixture revision, claim IDs, and boundary IDs as the web page. A materialized derivative is now labeled as a draft with a real repository path; an absent derivative remains planned with no invented path or review date. Existing public artifacts such as Course Echo’s lab and film are bound explicitly rather than represented as future work.

This structure makes additional collateral safer to produce. A presentation can shorten a story without strengthening its evidence. A film can dramatize a changed fact without implying an executed outcome. An interactive can expose an affected region and alternative branches without pretending to discover an unlimited optimum. A correction can mark dependent surfaces stale without silently rewriting what readers previously saw.

How to verify this release

1. Open the Scenario index and confirm that it contains eleven addressable Scenario cards.
2. Open each new detail page and inspect its semantic contract, eight-beat trace, story, capability ledger, authority chain, proof state, evidence boundary, and declared companions.
3. Compare Course Echo’s product-demonstration language with the illustrative evidence language on the other three entries. Confirm that the publication experience is not described as Grid-runtime evidence.
4. Open the Change One Fact lab, change the authored transport value, and confirm that the model exposes its arithmetic, source revision, affected views, branch conditions, and human-authority stopping point.
5. Follow a related Use Case or Insight and confirm that it links back to the Scenario that declared the relationship.
6. Confirm that the public site continues to use the repository’s existing Vercel project and grids365.fyi routes, with no secondary host or mirror.

Evidence, availability, and remaining proof

Public means these four editorial records, their links, and this permanent release record are included in the Grid FYI static production build. It does not mean a represented organization uses Grid, a domain workflow is integrated, an AI has decision authority, an instruction was executed, or an outcome occurred.

The new entries deliberately expose what remains. Intended-environment evaluations still require controlled sources, identities, permissions, mappings, policy and authority assignments, negative cases, security and accessibility review, external-system contracts, evidence retention, and acceptance thresholds. Research, public-warning, eligibility, defense, and other consequential uses require qualified domain owners. The release supplies better questions, reusable evidence boundaries, and inspectable authored contracts—not permission to skip those responsibilities.