Skip to content

chore(beads): apply corrective evidence contracts#2829

Merged
Sinity merged 1 commit into
masterfrom
feature/chore/corrective-beads-contracts
Jul 13, 2026
Merged

chore(beads): apply corrective evidence contracts#2829
Sinity merged 1 commit into
masterfrom
feature/chore/corrective-beads-contracts

Conversation

@Sinity

@Sinity Sinity commented Jul 13, 2026

Copy link
Copy Markdown
Owner

Summary

Apply the three-round corrective synthesis directly to the durable Beads graph. Fifty-seven owning Beads now stand alone around the five-axis unification test plus durability veto, honest result/evaluation/privacy contracts, typed domain boundaries, governed judgment/adaptation, external proof wedges, and contention-aware execution.

Problem

The analysis and continuity programs had converged across several design and adversarial-review sessions, but the durable graph still required chat context. It retained competing identities (MeasureSpec, cohort-pair experiments, claims ledger language), conflated epistemic axes (population/exactness, tag affinity/membership/confidence), under-specified authority boundaries, observational causal language, and a global lane cap that did not model actual contention classes.

Solution

  • Make rxdo the standalone program contract with identity/lifecycle/authority/access/durability unification tests, delivery phases, consumer-proof classes, and falsification receipts.
  • Bind queries to a language-wide protocol version plus component refs; make durability promotion-driven and evaluation-world receipts authoritative.
  • Separate enumeration, frame coverage, and measurement authority; wire coverage, uncertainty, rendering, privacy, retention, and excision owners without hard-blocking the P1 identity slice on the entire excision program.
  • Replace scalar tags with membership/affinity/confidence axes; embed EventOrderSpec in patterns; use horizon-aware unresolved inactivity.
  • Add the disclosure/instruction firewall with AssertionKind.POLICY; retain typed experiments as versioned assertion payloads until two consumers stabilize the lifecycle.
  • Preserve partial-order judgment semantics, two improvement-loop pilots, and a candidate-only curriculum experiment.
  • Correct D3/D8/D9 claim language; make bby.15 the named verified cold-reader export and 3tl.16 a claims view, not a second ledger.
  • Replace the global lane cap with contention-class admission plus a four-heavy-lane backstop; add collision-aware frontier and migration keys.
  • Fill execution-grade designs/ACs for the directly consumed pattern and rigor children.

The union export also preserves concurrent lane notes on polylogue-1dk1, polylogue-9e5.8.5, and polylogue-lkrc; those rows were not authored by this pass.

Verification

  • Contract audit: 57 corrective owners, 11 priority moves, 12 expected graph edges, and no broad privacy/frame hard blockers.
  • bd dep cycles — no dependency cycles.
  • .agent/scripts/bd-graph-lint — zero cycles/duplicate-label/priority-inversion findings; 27 missing-AC records remain, all proven present in the pre-change snapshot and outside this program.
  • JSONL/base audit — exactly 60 changed rows: 57 in-scope plus 3 preserved concurrent lane notes; no unexpected additions/deletions.
  • devtools verify --quick — all 15 checks passed in 55.5s.
  • Pre-push devtools verify --quick — all 15 checks passed in 52.0s.

Primary refs: polylogue-rxdo, polylogue-cpf, polylogue-ei94, polylogue-hg8n.

Problem:
The analysis, judgment, adaptation, and adoption programs had converged
across several reviews, but their durable Beads still contained competing
identities, ambiguous epistemic claims, and chat-dependent decisions.

What changed:
Make 57 owning Beads standalone around the five-axis unification test plus
durability veto, frame/evaluation/privacy contracts, authority firewall,
partial-order judgments, typed experiment assertions, two loop pilots,
honest demo claims, cold-reader export, and contention-class admission.
Preserve three concurrent lane notes through the live-database union export.

Verification:
- Contract audit: 57 corrective owners, 11 priority moves, 12 graph edges
- bd dep cycles: no dependency cycles
- bd-graph-lint: only 27 pre-existing missing-AC records; none introduced
- devtools verify --quick: all 15 checks passed in 55.5s

Co-Authored-By: OpenAI Codex <noreply@openai.com>
@coderabbitai

coderabbitai Bot commented Jul 13, 2026

Copy link
Copy Markdown

Important

Review skipped

Review was skipped as selected files did not have any reviewable changes.

💤 Files selected but had no reviewable changes (1)
  • .beads/issues.jsonl
⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: ASSERTIVE

Plan: Pro

Run ID: 0022a692-3161-424e-8b82-abadebac3fd1

📥 Commits

Reviewing files that changed from the base of the PR and between abad9b6 and 9a65319.

📒 Files selected for processing (1)
  • .beads/issues.jsonl

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

  • 🔍 Trigger review
✨ Finishing Touches
🧪 Generate unit tests (beta)
  • Create PR with unit tests
  • Commit unit tests in branch feature/chore/corrective-beads-contracts

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

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

@Sinity

Sinity commented Jul 13, 2026

Copy link
Copy Markdown
Owner Author

Publish-boundary triage:

  • Every GitHub Actions failure completed with zero steps. The check annotation is: "The job was not started because your account is locked due to a billing issue." This affects CI, CodeQL, Nix, and PR State Guard uniformly; it is external infrastructure, not a substantive gate result.
  • CodeRabbit skipped review because the only changed file is Beads JSONL; no findings.
  • GitGuardian passed; no inline review comments.
  • Local contract audit, dependency-cycle check, JSONL row audit, and two independent devtools verify --quick runs passed. The graph lint's 27 residual missing-AC records were present in the pre-change snapshot; this PR introduces none.

Admin squash proceeds under the repository's documented local-gate substitute for this account-wide outage.

@Sinity
Sinity merged commit 6116891 into master Jul 13, 2026
3 of 13 checks passed
@Sinity
Sinity deleted the feature/chore/corrective-beads-contracts branch July 13, 2026 06:05
@chatgpt-codex-connector

Copy link
Copy Markdown

💡 Codex Review

https://github.com/Sinity/polylogue/blob/9a653195bbefd1bb4c09169b4547328b0ae9783d/.beads/issues.jsonl#L610
P2 Badge Make stc a blocking dependency

This new contract says the analysis consumes ExperimentDefinition assertions from polylogue-stc and its acceptance requires a stc two-arm fixture, but the new edge is type":"related" with dependency_count":0. In Beads, only blocks edges contribute to dependency_count/ready filtering, so polylogue-rxdo.9.10 can be scheduled before the experiment-definition producer exists; change this edge to a blocking dependency and update the count so agents do not implement the projection against a missing or unstable stc contract.

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".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant