Readiness Status

This page keeps the public landing page clean while preserving the exact current truth for reviewers, agents, and release work. Garnet can be tried today; distribution, provider-backed assist, and backend claims remain evidence-gated.

Tracked implementation
87/87

Current implementation-plan slices are verified in the tracked ledger.

MIT/productization
92.3%

Broader productization remains active-partial and separate from tracked slice completion.

Converter truth
4 active

Rust, Ruby, Python, and Go are deterministic migration-assistant lanes.

Converter Taxonomy

GroupLanguages / targetsClaim boundary
Active conversion Rust, Ruby, Python, Go Deterministic migration assistant with sandbox, lineage, metrics, and migrate_todo evidence.
Advisory planning JavaScript, TypeScript, Swift, Java, C, C++, C#, Perl, Kotlin, Shell, SQL, Other Risk inventory, Garnet-aware context, advisory bundle, review, and handoff only.
Native boundary C, C++, Objective-C, Assembly, CUDA, platform-specific code Prefer native modules or FFI wrapped by Garnet CapCaps, memory declarations, lineage, and sandbox policy.
Backend lowering Wasm, LLVM-style native targets, native package toolchains Planned architecture only until compiler backend, tests, dogfood evidence, and benchmarks land.

Converter fit is decided by parser certainty and memory/capability preservation:

Editor/LSP Adoption

S16 builds on the S1 editor surface with a trivia-preserving CST, document/workspace symbols, CST-precise rename, rules-based quick fixes, and semantic tokens proven through the stdio protocol smoke.

SurfaceStatusRemaining evidence
garnet-lsp/ verified Diagnostics, hover, go-to-definition, document/workspace symbols, rename, quick fixes, and semantic tokens must stay green.
editors/vscode/ verified Local VSIX packaging, release-backed VSIX installation, diagnostic smoke, and the S16 stdio protocol smoke are the current evidence surface.
Parser CST layer verified CST round-trip tests reconstruct the parser-crate and workspace example corpus byte-identically; incremental and error-recovery parsing remain future work.

Formatter Baseline

S4 is now a merged, honest formatter baseline: deterministic whitespace, line-ending, and terminal-newline normalization enforced by integration tests. It is useful today, but it is not yet a full semantic pretty-printer.

SurfaceStatusBoundary
garnet fmt --stdout verified Two-pass idempotency is enforced on the canonical example corpus.
Semantic formatting deferred Spacing, alignment, import sorting, comment preservation, malformed-source recovery, and workspace-wide formatting wait for a trivia-preserving CST.

LLM Advisory Pipeline

Provider-backed conversion is not active. The feasibility reporter now exposes ten provider options as advisory-only evaluation candidates, and Garnet Studio can preserve that registry as a Desktop dogfood evidence bundle with source omitted by default and human approval required. The approved pipeline is:

source classifier -> risk inventory -> Garnet context -> advisory plan -> review handoff -> human-approved candidate -> garnet check/test/dogfood

Mac-Side Continuation

LaneStatusMac-actionable?Boundary
Reusable dogfood readiness skill Published Yes Navigata1/dogfood-readiness is a portable evidence gate, not a replacement for project CI or human release review.
Unsigned/local Garnet Studio quality Active-partial Yes Local UX, bundled evidence, DMG smoke, and app dogfood can improve before signing.
Studio Continuation Pulse Active Yes The Release panel can run the repo-native Mac-side continuation reporter from source or packaged resources while keeping Apple Developer ID and Windows/Linux gates separate.
Studio Deck Outline Active Yes The Release panel can create a manifested MIT deck-outline bundle from current evidence while keeping presentation planning separate from final acceptance.
MIT Deck Preview Active Yes The repo can render and browser-smoke the HTML deck preview in headless Chrome, checking desktop/mobile layout, screenshot evidence, readiness metrics, speaker notes, blocked gates, and forbidden claims while keeping human/aesthetic approval separate.
Studio Deck Preview Active Yes The Release panel can create a manifested HTML/JSON/outline deck-preview bundle, and copied-DMG app smoke now verifies that generated manifest through the app binary with a preserved checksum log while keeping final deck approval separate.
Website, status, and presentation Active-partial Yes Landing/status copy, demos, proof links, and repo-native MIT deck outline/preview artifacts can improve while preserving readiness truth.
Converter advisory quality Active-partial Yes Advisory plans, reviews, and handoff packets can improve without provider calls or active LLM conversion claims.
Proof, benchmark, and empirical evidence Active-partial Yes Add one falsifiable local evidence slice at a time.
Apple Developer ID notarization Blocked-external No Requires account-holder identity verification, Developer ID certificate, and notary profile before any notarization claim.
Windows/Linux Studio Windows active-partial No The Tauri v2 shell scaffold now builds on Windows, produces an unsigned NSIS bundle, writes smoke evidence, exposes the repo-native v0.5 readiness reporters, has verified x64 clean-VM installer proof, records WSL package/window portability evidence, verifies Studio Domain Proof Matrix shell output, and verifies Release / Readiness shell reporter output. Clean/non-WSL Linux desktop GUI launch, the live Release / Readiness GUI screenshot, signing, MSI, winget, and Windows ARM64 proof remain separate gates.

Open Product Gates

GateStatusNext proof
Apple Developer ID notarization Credential-gated Developer ID Application identity, notarytool profile, signed DMG, notarization, stapling, clean-machine Gatekeeper proof.
Windows/Linux Studio Windows source-build active, WSL portability active, Linux desktop open Preserve the clean Windows VM and Studio domain-shell bundles; next record clean/non-WSL Linux desktop GUI launch proof, then signed MSI/winget and Windows ARM64 gates.
Mobile distribution Future lane Separate iOS/Android product decision, app-store account setup, device tests, and release gates.
Provider-backed LLM assist Inactive Provider adapter, privacy controls, model evals, source inclusion approval, and dogfood readiness gate.
Native backend, proof, empirics Active-partial / planned Compiler backend evidence, mechanized proof work, benchmarks, and empirical validation.

Machine-Readable Sources