Repository navigation
Finalize portable grand catalog with independent Codex and Claude convergence - #55
Closed
seathatflowsinourveins wants to merge 26 commits into
Closed
seathatflowsinourveins wants to merge 26 commits into
seathatflowsinourveins wants to merge 26 commits into
Conversation
seathatflowsinourveins
marked this pull request as ready for review
September 21, 2026 23:07
…ent and record the workflow review Adds the sanitized 70-row report (47 asserted functional rows, 10 readiness, 3 version/help, 10 without local execution), 14 reviewed dashboard screenshots, a compact summary, projection provenance with names, sizes and hashes of all 266 privately retained streams, and a receipt whose embedded artifacts are a small declared subset. A check that does not invoke its component is labelled a local integration check and never counts toward a row. The page carries a Content-Security-Policy with a computed script hash. A new hash-listed unit test, run by the required validate job, enforces the bundle's invariants (five negative controls caught locally). GitHub workflow review, 2026-09-21: zizmor 1.30.1 regular persona 0 findings, attested actionlint 1.7.12 exit 0 on all nine workflows, all seven action pins resolve to their named and latest releases. Exact release comments added in two workflows; the same edit was reverted in three whose bytes are hash-bound (two frozen into recovery plans, where the repository's tests failed, and one recorded in four execution receipts), leaving six comments for the next functional change. Stricter-persona findings are retained with reasons. Reviewed before publication by a native Codex review and two Opus evidence reviewers; supported findings fixed (two rows no longer count local checks as component evidence; repository deep links; exact counts and bindings). Local CI-equivalent sequence: zizmor, actionlint, five validators, explorer rebuild check and 759 unit tests pass. Co-Authored-By: Claude Fable 5.1 <[email protected]>
…ents An independent evidence reviewer checked the branch at 34e2959: all 14 public screenshots and every new text file passed its privacy review, and all class, hash, tier and lifetime numbers recomputed exactly. It found that the guide said 25 rows have no audit record where 26 do (uv is listed as not contestable), that one anomaly sentence inverted the recorded Grafana defect and dropped its unexplained part, and that nothing bound the narrative counts in the guide, the explorer entry and the receipt claim to the results. Both statements are corrected and a tenth test now recomputes those counts. Local CI-equivalent sequence passes again: zizmor, actionlint, five validators, explorer rebuild check and 760 unit tests. Co-Authored-By: Claude Fable 5.1 <[email protected]>
seathatflowsinourveins
changed the title
The source template now presents a static link to index.html and keeps catalog controls hidden until initialization succeeds. Missing or corrupt data and render failures retain a useful recovery screen without redirects or background fetches. Verification: 40 ecosystem tests passed, including real startup-script execution with missing, raw, malformed and renderer-failure inputs. Native agent-browser followed template to a separately rebuilt index, observed 20 comparison layers, no external requests, no normal console errors and 390px mobile containment; injected corrupt data recovered visibly. Publication validation still reports expected hashes for these four changed files; the coordinator owns regeneration and hash refresh. Generated index was not changed here.
seathatflowsinourveins
changed the title
…d lean agent routing examples/claude-native now carries byte-identical copies of the deployed agent-lab files: three agents (source-scout Sonnet/medium for extraction and acceptance commands; evidence-reviewer Opus/high and isolated-builder Sonnet/medium with named MCP tools behind ToolSearch), the review-changes workflow with an independent recheck stage, readiness-audit, child-usage.mjs for per-child requested/resolved model and provider usage, the Codex bridge, and the local suites (static contract, mutation harness, synthetic usage rows, receipt bindings) driven by a sibling contract.config.json. ultracode.settings.json moves from the small to the unrestricted size guideline with a task sizing rule. examples/codex-native adds the matching Codex agent definitions and the [agents] config snippet. Evidence: receipt ultracode-token-routing-20260921 (first prompt 42,396 -> 12,164 -> 8,048 tokens on one identical task; eight native reviews, a readiness audit, a builder run and a Codex cross-family review as sanitized usage receipts; Haiku not routed after an Opus-verified 9/14 vs 14/14 trial) and receipt portable-claude-native-qualification-20260921 (the portable set adopted into a scratch project and run natively under the full host profile with a client-scoped observation that reconciles the Sonnet children exactly; the project-only profile is recorded as unqualified: provider errors and a permission-layer refusal of the Workflow tool). Decision lean-workflow-child-routing records the routing, the sizing rule and the comparisons that would overturn them; recipes gain the cooperation lanes and the observability README the per-model query. Co-Authored-By: Claude Fable 5.1 <[email protected]>
seathatflowsinourveins
changed the title
Codex reviewed 73f4334 against fb04bf7 (five findings, stored as codex-pr-review.json in the qualification bundle): the two new receipts and artifact bundles now publish at the current revision instead of the immutable base; the decision record keeps the project-only boundary in its scope and lifecycle; the observability README scopes a run by the client_scope label rather than the instance label; the recipe states inventory and recheck coverage separately; the Codex agent documents say that sandbox_mode is not set and the reviewer's no-edit rule is a prompt instruction. Co-Authored-By: Claude Fable 5.1 <[email protected]>
The portable review-changes workflow reviewed 73f4334 headlessly inside this checkout with its agents supplied inline (stored as claude-pr-review.json in the qualification bundle): the guide is regenerated from the current deployed record so its checks, owned files and the sync sentence match the shipped copies; the decision limitations carry the project-only boundary; the recipe names both readiness audits; the exporter coverage figure reads 10% in the guide and the receipt alike. Co-Authored-By: Claude Fable 5.1 <[email protected]>
seathatflowsinourveins
changed the title
This was referenced Sep 22, 2026
seathatflowsinourveins
added a commit
that referenced
this pull request
Sep 22, 2026
… union The union kept #57's corrected count in docs/full-stack-convergence.md but #55's stale 46 in catalogs/foundation/README.md; decisions.json holds 47 at every parent. Explorer regenerated and rehashed. Found by the independent evidence review of the integration. Co-Authored-By: Claude Fable 5.1 <[email protected]>
seathatflowsinourveins
added a commit
that referenced
this pull request
Sep 22, 2026
…e corrected default/override explanation PR #55's paragraph called min(16, CPUs − 2) the runtime ceiling; PR #57 had corrected it to the client default that the setting overrides (1–256). Found by the bounded Codex cross-family review of the integration. Co-Authored-By: Claude Fable 5.1 <[email protected]>
seathatflowsinourveins
added a commit
that referenced
this pull request
Sep 22, 2026
Merges the four open PRs (#55 with #56, #57, #58) into main with unioned shared indexes; validators, unit tests, CI, an Opus evidence review and a bounded Codex cross-family review passed on the branch. Co-Authored-By: Claude Fable 5.1 <[email protected]>
Collaborator
Author
seathatflowsinourveins
deleted the
codex/landscape-evidence-finalization-20260921
branch
September 25, 2026 18:47
5 tasks done
Merged
5 tasks done
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The catalog now gives a portable, evidence-led setup baseline across 16 foundation and four US-equities layers, with a working offline HTML handbook. The canonical index contains 843 repository identities; 208 candidate-layer rows compare 158 repositories. The 68 versioned components retain their pins, with separate shared-skill records and a locked research SDK.
Join independent Codex discovery with the live Claude Ultracode discovery, coverage critique and later screening repair. Preserve the original reports and failed attempts. Record 514 candidate-group rows, 441 distinct Claude discovery/refuter identities, all 159 later screening repairs and all 125 recommendation dispositions. Independent audits corrected screening arithmetic, source attribution, scope and misleading exclusions; Grype's unexecuted scanner card is conditional. Stars and reviewer agreement are discovery evidence, not a universal quality ranking.
The handbook includes per-layer choices, alternatives, reasons and overturn conditions; native Codex/Claude worker practice; TypeSafe and selected official skills; the bounded DeerFlow result; ai-memory versus Basic Memory lifecycle evidence; and explicit quality gates for research applications, retrieval, memory and the Nautilus/LEAN destination. Integrate the execution and native-workflow evidence from #56 and #57 without changing installed defaults.
The live Claude owner's post-publication review identified nine ambiguous summary labels. All 20 coordinator labels and handbook statuses now exactly mirror canonical layer decisions, with explicit pointers and a CI guard. Preserve the peer return, its per-child usage and the independent coordinator adjudication; no candidate disposition or selected pin was promoted. Clarify Codex's three-thread example and Claude's separate eight-worker cap without changing either setting.
The September 22 follow-up now records corrected comparison progress in a compact public evidence export and offline handbook guide. Independent review checked 163 sealed hashes, 38 repaired references, seven exact result recomputations and ten synthetic publication-gate tests. Preserve the defective original script, refuted judgments and incomplete worker receipt provenance. Comparisons remain unresolved or blocked; durable memory's passing reviewer gate does not establish a winner. Record the returned SPY parity run as blocked instead of planned, without changing the Nautilus destination or any selected version.
Validation:
853ca44: Python suite 786 tests, 784 passed and two environment-dependent skips. This progress-only update adds ten passing synthetic gate contract tests; it does not claim a new full upstream run.The September 21 documentary snapshot is complete within its stated scope; the live follow-up comparison work has not returned joint completion. The original critic's false readiness verdict and subsequent corrections remain visible. Same-host clean-prefix and memory checks do not certify the new PC; its native sign-in, useful tasks, scoped state and recovery will collect fresh acceptance. Matched quality/cost comparisons, complete research-service operation, paid hosting and broker execution remain within their own gates.