Source identity
| Field | Recorded value |
|---|---|
| Source type | project Library file |
| Original record | 00_READ_ME_FIRST.txt |
| Date | Sep 9, 2026 |
| Path or archive member | MegaMind/00_READ_ME_FIRST.txt |
| Original SHA-256 | 67ed4cb265d59684d5dd75430124b9d1be1fcec015a9ba870695eae9cd74c661 |
| Reading-copy SHA-256 | 67ed4cb265d59684d5dd75430124b9d1be1fcec015a9ba870695eae9cd74c661 |
Recovered source transcript
Download complete source text ↓64 lines · 3.7 KB
SEFERIM — SHAREABLE BUNDLE
Prepared 2026-09-09. Seferim was NOT run, started, or executed at any point.
────────────────────────────────────────────────────────────────────
PROVENANCE
────────────────────────────────────────────────────────────────────
Original (untouched, still in place):
iCloud Drive / AGI_Data / seferim
122,596 files · 67,461 directories · 1.5 GB · 0 iCloud placeholders
Forensic archive (complete, byte-for-byte, immutable):
~/Desktop/SEFERIM_FORENSIC/SEFERIM_ORIGINAL_RECOVERY.zip
Built with: ditto -c -k --sequesterRsrc --keepParent
1,277,854,532 bytes · 122,596 files verified · integrity test passed
sha256 ab73829cdfa668dbbc9cec2639292781226e777297d37c975903ea5787b4b20e
Working copy (full, unmodified):
~/Desktop/Seferim — 122,596 files, verified equal to original
────────────────────────────────────────────────────────────────────
WHY THIS BUNDLE IS SMALL — AND WHY NOTHING IS MISSING
────────────────────────────────────────────────────────────────────
Every one of the 122,596 files was SHA-256 hashed. Result:
122,596 files -> only 55,125 unique contents -> 71.3 MB of real data
The bulk was duplication inside snapshots/. Those 67,452 files hold
exactly THREE unique contents:
sha256 adb05ccc757157e7... 17,903 bytes repeated 67,025 times
sha256 07fe9638cdc392b4... 38 bytes repeated 426 times
sha256 345871e5663af7c1... 4,063,236 bytes snapshots/.DS_Store (1)
A checkpoint loop rewrote the same state every ~2 seconds between
2026-01-05 and 2026-01-09. 1.3 GB of snapshots is ~4 MB of real data.
NOTHING IS OMITTED:
* All 132 root/state files are here in FULL — including every tiny
one. All 37 *.state files (_g16 / _l16 / _m16 / _k16 variants),
federation_state.json and "federation_state 2.json". No file was
skipped for being small.
* memory/ federation/ substrate/ agents/ — complete.
* knowledge/ — complete. All 55,012 files, each one genuinely
unique (verified by hash), 172 MB.
* snapshots/ — all three unique contents are present in full bytes
under snapshots_deduped/, named by their hash.
* MANIFEST_FULL.tsv lists ALL 122,596 original files with sha256,
size and mtime — so any duplicate can be reconstructed exactly.
* SNAPSHOTS_INDEX.tsv maps all 67,452 snapshot paths to their hash.
To rebuild any snapshot: look up its path in SNAPSHOTS_INDEX.tsv, take
the matching file from snapshots_deduped/, done. Byte-identical.
If you need the literal 67,451 directories, use the forensic archive.
────────────────────────────────────────────────────────────────────
LAYOUT
────────────────────────────────────────────────────────────────────
00_READ_ME_FIRST.txt this file
MANIFEST_FULL.tsv 122,596 rows: sha256, size, mtime, path
SNAPSHOTS_INDEX.tsv 67,452 rows: sha256, path
seferim/ complete tree minus snapshot duplicates
snapshots_deduped/ the 3 unique snapshot contents