[AdaWorldAPI]

Docs and probe definitions only. Nothing is implemented, no owner is claimed, and nothing should merge automatically. This is the operator-approved follow-up to the GKAT excerpt in #1670 and the co-research map in #1668.

Changes

d-lfa-lance-graph-flow-rubicon-assimilation-v1.md gets a new section, Amendment 2026-10-21: guarded-control verification (GKAT). Where it differs from the existing D-LFA-1 and D-LFA-2 text, the amendment replaces it.

• Guard projection. A K3 predicate becomes a GKAT test only through a declared projection:

  • WHERE-style passes only TRUE ({T});
  • CHECK-style also passes UNKNOWN ({T,U}).

  The two are never interchanged. Commit authorization never inherits CHECK semantics.

• D-LFA-1. One structured Rubicon reference flow and one optimized variant.

  • Equivalence: they must be GKAT-equivalent, with the same action alphabet, declared guards and projections.
  • Separate trace check: no RequestCommit without a valid revision and MUL authorization.
  • RequestCommit is an intent, not a Lance write.
  • Disable runs: each must fail its check —
    • reversing a guard's polarity;
    • moving RequestCommit ahead of its guard;
    • introducing an unguarded nondeterministic choice.
  • GoTo and GoBack are refused, not translated.
  • WaitForInput is modelled as an observation of external state at resume.
  • Snapshot and codebook generation validation remains a separate runtime gate.

• D-LFA-2. Integration with what already exists:

  • the graph-flow driver;
  • thought-git verbs;
  • Kanban/Rubicon transitions;
  • authorized action proposals.

  Replay results and receipts are compared against the from-scratch oracle. No second scheduler, evaluator, state store or commit owner.

• D-GKAT-0. An independent, unclaimed fixture probe:

  • G1: guarded factoring and equivalence;
  • G2: genuine nondeterminism is refused;
  • G5: an explicitly bounded set of Boolean atoms.

  No production runtime, dependency or executor.

• Unchanged. D-RFT-2. D-RFT-4 keeps versioned thought binding and frontier behaviour; guarded workflow verification is not part of it.

Board.

• The co-research map gains two probe rows: D-GKAT-0 and P-GKAT-RUBICON.
• STATUS_BOARD:
  • D-RFT-0 is marked shipped (D-RFT-0 / D-QSP-0: Kleene form of set normalization (SQL NULL), PostgreSQL differential #1668);
  • new rows for D-LFA-1/2 and D-GKAT-0.
• Both indexes regenerated: entries --check passes; the supersession index is unchanged.

---

Generated by Claude Code

[coderabbitai]

Important

• 🔍 Trigger review

This repository does not receive automatic reviews because it has fewer than 10 stars.

⚙️ Run configuration

• Configuration used: Organization UI
• Review profile: CHILL
• Plan: Advanced
• Run ID: fc44dcea-27de-43f2-81c5-ab99904ed179

• Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

Warning

Billing warning: we have not been able to collect payment for this subscription for more than 72 hours. Please update the payment method or pay any pending invoices in Billing to avoid service interruption.

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	🔄 Running since 2026-10-21T19:20:24.092790Z	0eccbce	Draft marked ready

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.