Repository navigation
Public research shop: pinned views, one catalog, and restricted result intake - #145
Conversation
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
|
Source-exposed OpenAI engineering review receipt for Root validation: 85 Python controls, 16 frontend controls, complete 2,138-row interaction harness, exact public-source readback, generated-export comparison, and local navigation/landing checks passed. The four hosted checks on this exact head now all pass (verify, navigation, public-shop, intake-controls). Public tasks are #141–#144 and #146. Browser visual validation remains unavailable in this environment; settings are not claimed activated. A live trusted intake run will be checked after the workflow lands. |
A stranger can now follow one entry point to scoped status, full texts, a pinned coefficient explorer, and a contribution lane. This PR implements the owner's public-shop plan in main; Math- remains the proof vault.
Changes:
Validation: 85 Python controls, 16 frontend negative controls, full 2,138-row interaction harness, live pinned-source readback, export regeneration, navigation and landing checks passed. Nonauthor source-exposed engineering reviews found and repaired a source-symlink hole, skipped guard ambiguity, credential-shaped paths, and displayed-identity/URL mismatch; zero organizational-independence credit. Notebook cells and altered-input refusals verified. Browser visual validation could not run because the cloud browser cannot reach the local server.
Deployment limits: Pages, profile pins, Project board and branch protection require signed-in GitHub settings and are not claimed installed. A workflow file/check name is not by itself enforced protection; reviewers must verify the trusted guard run and exact head. A live fork intake run remains to verify after deployment. Secret scanning is explicitly limited, not a safety guarantee.
Scientific effect: NONE. No proof body, ACCEPT/AMEND row, lemma_closed, prize or landing claim is changed; Math #60/#64/#69 and main #122/#128 remain unmerged by this work.