storage intelligencelocalscan: setup-first CLI/app flowread-only
plan: cleanup-plan.json approvalsgated
clean: terminal-only dry runexplicit
Aark Optimize
Local-first Mac storage intelligence with explainable reports and approval-gated cleanup.
Implemented as Aark and the mac-storage-optimizer package: a Python CLI and local dashboard for technical Mac users to scan storage, generate deterministic reports, review cleanup plans, and clean only explicitly approved generated artifacts.
Stack
Python 3.11+ / macOS / local files / static dashboard assets / pytest / GitHub Actions
local-firstmacOS storagedeterministic reportsapproval-gated cleanupread-only dashboard
What Exists Now
- Scans selected local roots and optional local source groups into Markdown and deterministic JSON reports.
- Generates cleanup-plan.md and cleanup-plan.json with stable action IDs and approved: false by default.
- Serves setup-first app mode, demo mode, and a read-only local dashboard from packaged static assets.
Technical Signals
- pyproject.toml exposes the mac-storage-optimizer console script and classifies the package as alpha/private beta.
- mac_storage_optimizer/cli.py dispatches app, dashboard, demo, compare, sources, approve, and clean workflows.
- 43 regression test functions cover scanner output, storage decisions, source discovery, dashboard safety, approval tokens, cleanup guards, demo mode, and launcher contracts.
Why It Matters
Mac storage pressure usually mixes safe generated artifacts with user data, project state, and cloud sync folders. Aark Optimize makes the evidence visible before cleanup and keeps destructive action separated behind explicit approval.
Known Limits
- Private technical beta, not a public downloadable app, autonomous cleanup service, or cloud sync product.
- The Finder .app is a lightweight launcher, not a signed or notarized macOS application.
- Cloud support is local sync-folder mapping only; it does not call iCloud, Google Drive, Dropbox, or GitHub APIs.
- Move/archive recommendations are review-only and do not verify destination capacity, checksums, or rollback.
Experimental testnet infrastructure hash: SHA-256 raw bytesstored
anchor: XRPL Testnet txguarded
verify: file or hash endpointpublic
LedgerSeal
XRPL Testnet proof infrastructure for document hash anchoring and independent verification.
A Next.js App Router application that hashes document bytes, anchors hashes to XRPL Testnet, stores LS-1 proof metadata in Postgres, and exposes org-scoped dashboard workflows plus public verification.
Stack
Next.js App Router / TypeScript / Prisma / Postgres / XRPL JS SDK / Tailwind / BullMQ/Redis plumbing
XRPL TestnetSHA-256proof bundlesorg-scoped APIpublic verification
What Exists Now
- POST /api/seal accepts multipart, octet-stream, text, or JSON content, requires x-org-id, hashes raw bytes, submits an XRPL Testnet transaction, and stores proof metadata.
- GET /api/seals and GET /api/seals/[id] expose org-scoped list/detail APIs with status, date, network, query, cursor, and limit handling.
- POST /api/seals/verify accepts file upload or SHA-256 hash input, rate limits public verification, checks hash match, and confirms the XRPL transaction.
Technical Signals
- Five API route files implement health, seal creation, seal listing, seal detail, and verification endpoints.
- App Router pages cover the landing, public verify screen, dashboard, seal upload, seal detail, and redirect compatibility paths.
- lib/xrpl-config.ts and lib/testnet-guardrails.ts reject non-testnet configuration and mainnet URLs.
Why It Matters
Operational records often need an integrity check that survives outside the app that created them. LedgerSeal keeps the stored record small: hash the bytes, anchor the hash, store proof metadata, and let internal or public verifiers check the result later.
Known Limits
- XRPL mainnet support is explicitly out of scope; the code enforces Testnet.
- BullMQ and Redis worker modules exist, but workers are not wired to a runtime entrypoint yet.
- Authentication is currently org header/API-key scoping, not full user accounts or role management.
- Automated test coverage is narrow today: one Node test file covers hash vectors.
- A proof confirms byte integrity and ledger anchoring, not authorship, content accuracy, or first existence.