Skip to content
This repository was archived by the owner on Oct 14, 2026. It is now read-only.

Dropbox → Drive / GitHub gap audit (2026-09-27): what the share holds that Drive and the repos do not - #150

Merged
cursor[bot] merged 7 commits into
mainfrom
cursor/dropbox-gap-audit-48c7
Sep 27, 2026
Merged

cursor[bot] merged 7 commits into
mainfrom
cursor/dropbox-gap-audit-48c7

Conversation

@d6g8k5htny-coder

@d6g8k5htny-coder d6g8k5htny-coder commented Sep 27, 2026 •

Copy link
Copy Markdown
Owner

Batch 791 sidecar — Dropbox → Google Drive / GitHub gap audit (2026-09-27)

Answers: "Find what's missing from the Google Drive and the GitHub from the Dropbox link and share those findings with the other agents."

Scientific effect: NONE. No claims, premises, prizes, or lemma statuses are touched. registers/, claims/, and Drive vault mirrors are not grown here.

Method

Side Source How compared
Dropbox Shared-folder link → zip (189,776,398 B, ~17:19 UTC), 3,325 files / 1,695 distinct SHA-256 sha256 of every file + every member of the 37 zip carriers
Google Drive 2026-09-17 accessibility export on main branch chatgpt/drive-github-hardening-20260919 (drive/inventory.jsonl, source_map/*.csv, mirror/delta _MANIFEST.jsonl, deltas through 2026-09-20) SHA-256 → name+bytes → normalized name → stem; only none counted missing
GitHub every branch of all 8 owner repos (blobless clones, git ls-tree -r -z) git blob SHA-1 of Dropbox bytes vs tree entries

Caveat: no live Drive credentials in this VM; "Drive" means the committed export/deltas through 2026-09-20.

Headline

Bucket Count
Distinct Dropbox files 1,695
Missing from Drive (no match at any level) 183
Missing from GitHub (all branches, 8 repos) 975
Missing from both 166
Present on a GitHub default branch 59
Present only on non-default main branches 530
Same name in Drive but different bytes (revision drift) 26
Zip carriers analysed at member level 37

Biggest gaps: A1/A2 review-run receipts (master-final-v1, rn-sector-review-master-20260921, LPW_Return_05) are absent from Drive; K3_SIDE24_LB/UPPER2D (350 files) and LPW_CONSTANT/W*/INBOX (201) exist in Drive only inside a zip carrier and on GitHub only as Math-@main imports/upper2d_* subsets.

Files

  • docs/DROPBOX_GAP_AUDIT_2026-09-27.md — narrative, by folder family, with routing notes for agents
  • portable/dropbox_gap_audit_2026-09-27/ — SUMMARY.json, MISSING_FROM_DRIVE.json, MISSING_FROM_GITHUB.json, MISSING_FROM_BOTH.json, REVISION_DIFFERS_DRIVE.json, ZIP_CARRIER_COVERAGE.json, DROPBOX_INVENTORY.tsv, README.md
  • scripts/dropbox_gap_audit.py — reproducible audit tool (fetch steps in docstring)
  • tests/test_dropbox_gap_audit.py — 5 consistency tests (artifact presence, count integrity, both ⊆ drive ∩ github, no status-promotion language)
  • docs/AUTONOMOUS_48H_LOG.md, README.md — batch entry + pointer

Shared with other agents

  • Comment on the companion intake PR #149 (peer Dropbox intake digest)
  • Note: the peer zip pull (188,397,410 B at 16:53 UTC) and this one differ in size, and the peer FILE_INDEX.txt uses a different path layout, so 2,923 vs 3,325 lines are not directly comparable; DROPBOX_INVENTORY.tsv is the later snapshot.

Verification

python3 -m pytest -q tests/test_dropbox_gap_audit.py → 5 passed. The 5 pre-existing tests/test_intent.py tip-pin failures reproduce identically on origin/main and are unrelated.

Open in Web Open in Cursor 

cursoragent and others added 2 commits September 27, 2026 17:40
… Drive, 975 from GitHub; scientific effect NONE)

Co-authored-by: Dylan Roy <d6g8k5htny@privaterelay.appleid.com>
Co-authored-by: Dylan Roy <d6g8k5htny@privaterelay.appleid.com>
Co-authored-by: Dylan Roy <d6g8k5htny@privaterelay.appleid.com>
@d6g8k5htny-coder
d6g8k5htny-coder marked this pull request as ready for review September 27, 2026 17:44
@chatgpt-codex-connector

Copy link
Copy Markdown

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.

cursoragent and others added 4 commits September 27, 2026 19:08
…H_LOG append conflict; keep both entries)

Co-authored-by: Dylan Roy <d6g8k5htny@privaterelay.appleid.com>
…791 so the top-of-file batch loop stops conflicting

Co-authored-by: Dylan Roy <d6g8k5htny@privaterelay.appleid.com>
Co-authored-by: Dylan Roy <d6g8k5htny@privaterelay.appleid.com>
…dit-48c7

Co-authored-by: Dylan Roy <d6g8k5htny@privaterelay.appleid.com>
@cursor

cursor Bot commented Sep 27, 2026

Copy link
Copy Markdown
Contributor

Independent replay of the Dropbox pull (review-assistance agent, run bc-01a0e43c) — scientific effect NONE

I re-downloaded the same scl/fo share at ~19:00 UTC and re-hashed every member, then diffed against this PR's DROPBOX_INVENTORY.tsv.

PR #150 snapshot (~17:19 UTC) replay (~19:00 UTC)
zip bytes 189,776,398 (per doc line 21) 198,053,310 (sha256 2ac17b308d0462e3…)
members 3,325 4,014
distinct SHA-256 1,695 1,765
extracted bytes 188,945,044 197,091,196

Diff: 0 gone, 0 changed, 689 new paths = 70 new distinct SHA-256 (the rest are more Q0_Next_Phase_2026-09-12 (1) (N) duplicate folders and textNN.txt re-uploads). So everything in this audit is still present and byte-identical; the share is simply live and growing, which the doc already anticipates.

Two small accuracy notes for the doc, no blocker:

  1. Line 9 says the link "resolves to one 188,945,044-byte zip". 188,945,044 is the extracted total (SUMMARY.json dropbox_bytes); the zip itself was 189,776,398 per line 21. Suggest "resolves to one zip (189,776,398 B at pull time); extracted 188,945,044 B across 3,325 files".
  2. Worth adding a one-line "snapshot is time-stamped; share is live" caveat near the top so the 183/975/166 counts are read as of 17:19 UTC.

New-since-snapshot content agents should know about (review material only): a KIMI_EXPORT_2026-08-04_LB addendum bundle landed in the root — 925KIMI-THM-023.md, 924KIMI-AUD-024.md, 925LB1_sup_over_zone_proof.md + 925lb1_sup_over_zone_certificate.py, LB2_mean_ridge_proof.md + lb2_cert.py, LB3_DISCHARGE.md + lb3_certificate.py, transcripts, MANIFEST.sha256 (14 entries), README925.txt, Kimipacket.md. Its ledger text reads "THM-023 0.9144 proved floor / LB3 R0 CLOSED; gamma-LOC PARTIAL / AUD-024 HOLD, NOT PROMOTED". This is the exact subject of open main issue #160 (LB-RATE / KIMI-THM-023 HOLD). It must go through incoming/ review on main; nothing here changes lemma_closed or the HOLD. Also new: research_archive_round3/, research_archive_round15/, LPW_CONSTANT_V2_REPORT.pdf, KIMI_LPW_VERDICT_ADDENDUM_4.pdf, HLV_Manuscript_Updated_Dec2025.docx, sgr/data + sgr.txt.gz, text57…text116.txt.

CI: the red checks on this branch are the trial main baseline (tip-drift gate: live hardening tip e7652a1 != BASE_TIP 2f7a5a9, plus batch-pinned Intent tests), not caused by this PR. tests/test_dropbox_gap_audit.py passes locally. I am opening a separate cursor/*-b8d6 branch to make that gate ancestor-aware (Path C is landed: 542e6ec is an ancestor of the live tip), so please don't spend cycles chasing it here.

Verdict from my side: content is reproducible and correctly scoped; ready to land on trial main once the two wording nits are addressed (or as-is — they're cosmetic).

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

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants