What is built, what is next, and where every claim came from
Stated plainly, including the parts that are not finished. A status page that only lists successes is not a status page.
Current status
| Component | Detail | Status |
|---|---|---|
| Karsa Quantum OS Web control plane · 12 modules |
Twelve modules simulate the full quantum operations command centre end to end, now including the HNDL Batch & Vault Lab and QKD Security Lab. Runs standalone from local state; the adapter is optional and never required. | Stable |
| Quantum systems simulator 4 systems |
Verified against a live run, 26 of 26 checks. Seven defects documented with exact line numbers and proposed fixes; source intentionally left unmodified pending review. Three would block any standard client. | Documented |
| Adapter HTTP + event stream bridge |
Live for task submission, resource inventory and monitoring across all four systems. Three connection modes degrade gracefully. Authentication and multi-user sessions are not yet implemented — localhost only. | Partial |
| PQC readiness scanner POSIX shell |
Shipping. Seventeen checks, nine scored, structural caps, evidence classes, stricter-profile support, machine-readable output, bilingual, consent-gated sharing, CI exit code. Windows parity is next. | Live |
| Qryptonite & toolkit 9 crates · WebAssembly |
ML-KEM-768 and ML-DSA-65 live in the toolkit and CLI; all 12 SLH-DSA parameter sets in the library, held out of the browser build while the backend is a release candidate. Known-answer vectors pass. Not FIPS 140-3 validated, not externally audited, pre-1.0. | Live · scoped |
| HNDL & QKD Security Labs Security extensions |
Both now live in the application. The HNDL lab sweeps the full asset inventory and generates a synthetic harvest snapshot; the QKD lab runs a real per-pulse BB84-family Monte Carlo with four Eve strategies, verified against analytic expectations. Neither is independently audited or claimed as production security. | Prototype live |
Scroll the table horizontally to see all columns.
Seven steps, ordered by value per unit of effort
Fix the simulator at source, then retire the shims
Apply the seven documented patches directly. The adapter's wire-level compatibility workarounds exist only because the defects do, and every shim left in place is a divergence between what the contract says and what the client sends.
HNDL risk engine — batch, not slider Built
Shipped as the HNDL Batch & Vault Lab: the single-asset slider extended to a full portfolio sweep with scenario modelling and a synthetic harvest-snapshot generator. This is what makes the control plane necessary rather than decorative.
Wire the Optimizer and Benchmark Lab to real submission
Both remain local-only because the simulator has no equivalent operation. Where one exists, dispatch through the adapter so benchmark campaigns and optimisation runs inherit the queue, the retry semantics and the evidence trail.
Synthetic HNDL laboratory Built — MVP
Generates weighted synthetic session records — asset, algorithm, capture date, classification, retention, projected decrypt-by year. Synthetic by design, metadata only. Full object-level scenarios remain a future enhancement; the production-key boundary is enforced in the product, not in a disclaimer.
Authentication and multi-user sessions in the adapter
The prerequisite for any shared deployment, and the gate between the single-host lab and an enterprise topology. Roles must become enforced scopes; secrets must move to a vault.
QKD BB84 MVP Built — exceeds MVP scope
Shipped as the QKD Security Lab: seeded channel model, four protocols, four Eve strategies, error estimation, sifting, and an asymptotic secret-key estimate — reproduced against analytic expectations in testing. What remains is validation against an independent QKD library or published reference vectors, before any physical device is attached.
Photonic and QPU integration
Only after the software models, validation vectors, identity and telemetry contracts are stable. Simulator-first stays the architecture; a real QPU backend can be added later without changing the contract, which is the whole point of having one.
Every claim, traced
Primary sources reviewed September 2026. Where a figure moves — and several of these move fast — the date it was read matters as much as the figure.
Origin Quantum
- Origin PilotOS 4.2 User Manual and API Reference, v3.4, 29 April 2026 — architecture, kernel modules, console, APIs, message push, type and error dictionaries, deployment
- Origin Pilot product pages and download guide — six core capabilities, deployment scenarios, editions
- Origin Wukong 180 specifications and the 2026 China machines comparison
Standards and government
- NIST FIPS 203 / 204 / 205 — ML-KEM, ML-DSA, SLH-DSA, finalised 13 August 2024
- NIST IR 8547 (initial public draft) — deprecation after 2030, disallowance after 2035
- CISA · NSA · NIST joint quantum-readiness fact sheet, August 2023 — the harvest-now-decrypt-later warning
- NSA QKD guidance and CNSA 2.0
- BSSN — National PQC Migration Roadmap Task Force, kick-off 25 June 2026
Cryptanalysis research
- Gidney (2025) — factoring RSA-2048 with under a million noisy qubits, a 20× reduction on the 2019 estimate
- Iceberg Quantum (2026) — the Pinnacle qLDPC architecture, under 100,000 physical qubits. Simulation and resource estimates, not an experimental demonstration
- Google Quantum AI et al. (2026) — ECDLP-256 under 500,000 qubits with zero-knowledge disclosure of the resource estimate
QKD security literature
- Photon-number splitting, Trojan horse, detector blinding and faked state, time-shift, dead-time, phase remapping and electromagnetic side channels
- Countermeasures: decoy states, optical isolation, power and temporal monitoring, detector self-testing, and measurement-device-independent QKD