Conversation
…evice The formats lived only in the readers, writers and the device layer. This writes them down with a status on every field: verified on a real card, only what the code does, or unknown. Verified on a v7 `.md` device (USB 0483:a341, firmware 3.2.3), read-only: the `.md` layout including two fields the parser skips (the USB vendor and product ids), the `.pi` index, the folder naming, the `ni` header and 44-byte stage node, `li`, the 12-byte `ri`/`si` entries, the 512-byte cipher boundary, the zero padding of the V3 cipher, the word-swapped AES key (openssl reproduces the library files from the card), the serial-derived `bt`, the BMP asset format, and the fact that three of the nine Luniistore packs on the card use a key that is not the device key. Unknowns are listed in the order they are worth attacking, with what was already tried against them so nobody repeats it. A section on versioning states what the formats offer for keeping several versions of one story: a version number nobody is known to read, and a UUID that is the only identity the device knows. No device data is committed; keys and serials are described, not reproduced. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
4 tasks
…pack and writes nothing to the card Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
…:a341 device Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
…d the USB reset on 0xf6 Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
…y shields STUdio packs from a Luniistore sync Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
… unknowns The deep firmware work is already done — o-daneel's Lunii.RE (v1/v2) and Lunii_v3.RE (v3) reverse the STM32 in Ghidra. Folding in what it settles: - The three opaque-bt Luniistore packs are not an anomaly: on v3, bt is a per-story AES key+IV ciphered with the device key, and resources use that per-story key. So they are genuine v3 packs read by a known path (decipher bt with the device key, then resources with the story key), not the mystery §10 called them. - .cfg is decoded, not unknown: nine (tag, value) pairs, meanings read from the firmware, matching this card. Replaces our guessed structure. - wifi.prefs is the device's own Wi-Fi credentials under the device key, not the app's. - Corrected an overclaim: .md 0x34 is not the USB vendor/product ids. The bytes equal them by coincidence, but Lunii.RE documents 830441A3 as a static signature; the field is left open rather than mislabelled. The unknowns section becomes "prior art, and what is still open", crediting the three repos and reordering what is left. STUdio needs none of the crypto; this is interoperability context and honest sourcing. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
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
1 participant
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.
What
FORMATS.md: the on-card layout (.md,.pi,.content/), the pack files (ni,li,ri,si,bt,nm, assets), both ciphers, and a section on what the formats offer for versioning a story. Every field carries one of three statuses: verified on a device, code only, unknown. Linked from the README next toTESTING.mdandFIELD-VALIDATION.md.How
A read-only session on a v7
.mddevice (USB0483:a341, firmware 3.2.3, 12 packs: 3 STUdio-made, 9 Luniistore). Nothing written to the card, no device data committed — keys and serials are described, never reproduced.Worth knowing
.md, word-swapped asAESCBCCipherdoes, andopenssl enc -aes-128-cbcon the first 512 bytes of a card file gives exactly the clear library file. The 512-byte boundary and the zero padding of the V3 cipher are confirmed on real files..mdv7 carries the USB vendor/product ids at 0x34/0x36, in the 14 bytes the parser skips. Two more unread 16-byte fields at 0x60 and 0x70 are documented as unknown.3.2.3→ 3.2). A3.10firmware would be misread. Documented as a gap, not fixed here.bton v7 is serial-derived and pack-independent on 9 of 12 packs — but three Luniistore packs carry an opaquebtand are ciphered with a key that is not the device key. What was tried against them is listed so nobody repeats it.ni— the failure mode a versioning feature would need to address. That feature is tracked separately.Docs only; CI unaffected.
🤖 Generated with Claude Code