# AutoForge ## Product vision The following describes the product vision, not implemented capabilities: AI-assisted automotive diagnostics and programming platform. Read and back up vehicle data or build AI-researched vehicle profiles. Simulate control modules and driving scenarios, replay telemetry, and test calibration changes before real-world programming. Diagnose faults, read/clear trouble codes, and program supported vehicle modules. ## Implementation status ### Implemented - Step-01 platform-neutral Core/Application foundation and regression coverage. - Step-01A single Windows-first .NET 10 MAUI Blazor Hybrid Client, UI and generic Components RCLs. Step-01A was accepted by the orchestrator on 2026-10-28. - Step-02 validated vehicle/profile/module metadata, strict schema-v1 JSON interchange, an in-memory revision catalog, and a synthetic profile UI. Step-02A corrected strict timestamp parsing and optional evidence applicability. Step-02 and Step-02A were accepted by the orchestrator on 2026-10-28. - Step-03 adds a provider-neutral diagnostic protocol/session path and a data-driven synthetic virtual-module provider. The native UI explicitly loads and selects a saved profile revision/provider, opens a separate simulation session, probes endpoints, reads module identity/DTC/static signal snapshots, clears only supported virtual DTC state after fresh empty readback verification, and exposes a bounded raw/decoded trace. Step-03A corrects strict wire presence/ ambiguity checks, complete reply validation and truthful trace outcomes. Step-03B preserves successful probe/service evidence separately from optional profile-verified identity and keeps expected/probed/identified counts distinct. Read/clear actions remain gated on verified identity. Step-03 with its corrections was accepted by the orchestrator. - Step-04 adds immutable versioned dynamics configurations, a fixed-step coupled synthetic engine/transmission/longitudinal model, an explicit `autoforge.simulation.dynamics` provider using protocol v2, configuration import/export, lifecycle/scenario controls, and source/unit/model-time-aware telemetry at `/simulation`. Step-04A corrects typed reference-validation failures, drivetrain/rev-limit consistency, monotonic phase/overall progress, terminal-failure telemetry/UI behavior, and single-owner cancellable pacing. Step-04B corrects side-effect-free Start/Resume rejection during owner cleanup. The static provider and protocol v1 remain separate. Step-04/04A/04B are accepted by the orchestrator. - Step-05 adds one configurable numeric/bar dashboard panel to both static diagnostics and dynamics workflows. It reads only identity-verified signals from the explicitly opened session, exposes source/quality/range/sample context, and keeps acquisition session-owned while gauges are hidden or removed. Step-05 with Step-05A/Step-05B was accepted by the orchestrator. - Step-06 adds a shared, persisted display preference with an Imperial (US) / Road first-use default and Metric/Imperial (US) plus Road/Engineering choices. Supported readouts convert only at presentation; SI source values, model calculations, configuration fingerprints and canonical dashboard definitions remain unchanged. See [display units](docs/live-dashboard.md#display-units). Focused correction Step-06A is accepted. - UI routes: `/` foundation, `/execution-targets` explicit metadata selection, `/vehicle-profiles` explicit profile import/revision selection/inspection/export, `/diagnostics` static synthetic diagnostics, `/simulation` dynamics, and a visible not-found route. - Capability declarations distinguish Supported, Unsupported, Unknown, Inaccessible and Restricted. Missing declarations are explicitly Unknown. Provenance and all profile facts are supplied metadata, not independently verified truth or permission. - Profile import is atomic and unknowns remain explicit. A selected profile is immutable, process-local metadata; it is not connected to a vehicle. - Publication/retrieval timestamps require a complete date/time with seconds and an explicit `Z` or signed `HH:mm` offset; partial dates/times and local-time defaults are rejected. A null optional module identity in evidence is no constraint; supplied values must match a known module identity and never fill unknown module data. See the schema guide for exact grammar and limits. ### Accepted recording/archive milestone Step-07 adds bounded in-memory recording of actual static-v1 and dynamic-v2 diagnostic operations, explicit source-sensitive local archive save/readback, strict schema-v1 validation, and shared recording/inspection controls in both host workflows. Step-07A corrects source-faithful empty/case-variant response round-trips, v2 archive validation, capture lifetime release, and visible storage failures. Step-07B removes the source-close versus callback-unregister lock cycle while preserving shared disposal completion and failure reporting. Step-07 with Step-07A/Step-07B was accepted on 2026-10-30. The three worker reports retain separate historical execution evidence. See the [Step-07 worker report](prompts/Step-07-Report.md) and [Step-07A corrective worker report](prompts/Step-07A-Report.md) and [Step-07B lock-safety corrective report](prompts/Step-07B-Report.md). ### Accepted recorded replay milestone Step-08 adds `/telemetry-replay`: explicit saved-capture loading without a profile catalogue, vehicle session or running model. One page-owned, cancellable playback clock projects exact recorded observation prefixes into the shared numeric/bar gauge editor. Navigation, partial-source acknowledgement, recorded provenance and runtime speed are explicit; source archives are never rewritten. See [replay usage and contracts](docs/telemetry-replay.md) and the [Step-08 worker report](prompts/Step-08-Report.md). Step-08 was accepted on 2026-10-30. ### Accepted synthetic authoring milestone Step-09 adds `/calibration`: one bounded, process-memory synthetic authoring workspace derived explicitly from a saved immutable dynamics configuration. Warm-idle scalar and fixed-axis torque proposals have exact source/head preconditions, immutable linear revisions and structured canonical diffs. These are Baseline/Proposed authoring values, not ECU backups, Tested/Approved calibrations or applied model/device changes. See [calibration authoring contracts](docs/calibration-authoring.md) and the [Step-09 worker report](prompts/Step-09-Report.md). Step-09 was accepted on 2026-10-30. ### Accepted isolated programming rehearsal milestone Step-10 adds an explicit isolated Simulation rehearsal within `/calibration`. Its real mutable logical image supports independent volatile backup, source/head/revision-bound single-use approval, atomic CAS, full readback, read-only reconciliation and separately approved original Restore. It never applies authoring values to a dynamics configuration or a physical target. See [rehearsal contracts](docs/calibration-rehearsal.md) and the [Step-10 worker report](prompts/Step-10-Report.md). Step-10A corrects close ownership while a new rehearsal factory return is pending: a newly admitted opening lifetime cannot reuse the previous close completion, and Close/disposal drains its actual owner before returning. The [corrective report](prompts/Step-10A-Report.md) records controlled pre/post behavior and native recheck. Corrected Step-10/10A was accepted on 2026-10-31; no programming scope or limits were added by the correction. ### Accepted deterministic preview milestone Step-11 adds `/runbooks`: strict manual schema-v1 JSON validation against an explicit current synthetic authoring source/head, complete template rules and exact read-only candidate/diff preview. Invalid input retains the last successful preview, and source changes label it Stale without rebasing. There is no model, retrieval, authoring Save, approval, target access or runbook execution. See [runbook schema/preview contracts](docs/runbooks-v1.md) and the [Step-11 worker report](prompts/Step-11-Report.md). Step-11 was accepted on 2026-10-31. ### Accepted local generation milestone Step-12 adds an explicit local-only Ollama adapter and matched synthetic knowledge/context in `/runbooks`. Model-generated content goes unchanged through the same deterministic validator with atomic source/prior-preview admission. There is no Save, approval, tool or target execution path. Controlled HTTP/UI/ lifetime tests and the complete build pass. Under the corrected owner-directed Step-12A handoff, the existing Ollama 0.40.13 now has a verified current cloud-disabled:true startup, preserving the exact installed `huihui_ai/qwen3-abliterated:8b-v2` digest/model store. Actual native original output produced exactly +600 rpm/-10 N*m with every other source value unchanged. Real pending conflicts, Cancel/drain, unmount/fresh-owner, manual/stale/units and source-isolation checks passed. A narrow Hybrid event/transport-dispatch fix keeps pending controls responsive; all 596 tests pass (594 retained plus two corrective cases). Step-12/12A was accepted on 2026-10-31. See [local generation contracts](docs/local-runbook-generation.md) and the [original Step-12 worker report](prompts/Step-12-Report.md), preserved alongside the [Step-12 completion report](prompts/Step-12-Completion-Report.md). The [previous Step-12A report](prompts/Step-12A-Report.md) remains historical; current evidence is in the [Step-12A execution report](prompts/Step-12A-Execution-Report.md). ### Verified synthetic comparison; pending manual review Step-13 adds explicit current-preview handoff into the existing pinned draft, separate Proposed Save and owning verified rehearsal-image admission. Evaluation projects only warm-idle/torque cells into a NEW immutable configuration and compares two fresh actual Simulation/v2 runs. Fixed acquisition is every five model steps plus completion, bounded at 3900 steps per run inside unchanged archive-v1 limits. Canonical measured metrics, proof/config/run/capture lineage, explicit per-capture save/readback, real RecordedReplay and cancellable/stale ownership are separate from proposal and programming evidence. See [comparison usage](docs/calibration-comparison.md) and [Step-13 worker evidence](prompts/Step-13-Report.md). Step-13 requires manual acceptance; Step-14 remains blocked. Step-13A corrects the remaining owner-release tail: Busy and the exact admitted drain stay reserved through cleanup, before atomic terminal completion. See [corrective execution evidence](prompts/Step-13A-Report.md). Step-13/13A remain pending manual acceptance; no model or programming scope is added. ### Planned Durable/full-device backups, broader model application, AI/retrieval/executors and real hardware operations require later approved assignments. Saved recordings are local file archives, not crash-durable live streaming. Replay does not resimulate or interpolate. Dynamics estimates are synthetic, not manufacturer prediction, vehicle compatibility, physical access, safe tuning advice or authorization. No capability is implemented merely because it appears in the product vision or roadmap. ## Architecture One Windows-first **AutoForge.Client** .NET 10 MAUI Blazor Hybrid executable hosts the application UI via `BlazorWebView`. **AutoForge.UI** owns all routes, navigation, layouts and feature presentation. **AutoForge.Components** owns vehicle-independent reusable controls. Application features run in-process in a modular monolith, and Core remains platform-neutral. There is no backend, browser-hosted WASM shell, Blazor Server host or second application executable. The application slices provide explicit target metadata resolution, vehicle-profile metadata import/list/select/export, dynamics configuration catalog/import/export, simulation-only diagnostic sessions, bounded session recording, archive inspection, offline recorded replay, synthetic calibration authoring, isolated simulated programming rehearsal, deterministic read-only runbook preview and a shared display-unit preference with a native local-preference adapter. The Client composes separate static and dynamics synthetic providers through provider-neutral factory contracts and a session-lifetime in-memory profile/ configuration catalog. Import never opens a session; metadata resolution never opens a session or starts a run. ```mermaid flowchart TD Client[MAUI Blazor Hybrid Client] --> UI[UI RCL] Client --> Simulation[Simulation provider] UI --> Components[Reusable Components RCL] UI --> App[Application feature slices] Client --> App App --> Diagnostics[Provider-neutral Diagnostics contracts and protocol] App --> Core[Platform-neutral Core models] Diagnostics --> Core Simulation --> Diagnostics Simulation --> Core UnitTests[Headless unit tests] --> Core UnitTests --> App UnitTests --> Diagnostics UnitTests --> Simulation IntegrationTests[Headless integration and rendered UI tests] --> UI IntegrationTests --> App IntegrationTests --> Core IntegrationTests --> Diagnostics IntegrationTests --> Simulation ``` Core and Application own their behaviors under feature-focused source folders: - `AutoForge.Application/Features/ExecutionTargets/SelectExecutionTarget` - `AutoForge.Application/Features/Vehicles/Profiles` - `AutoForge.Application/Features/Vehicles/Diagnostics` - `AutoForge.Application/Features/Vehicles/Simulation` - `AutoForge.Diagnostics` (provider-neutral protocol, transport/session contracts) - `AutoForge.Simulation` (data-driven virtual modules, fixed-step engine and transports) - `AutoForge.UI/Features/Vehicles/Profiles` - `AutoForge.UI/Features/Vehicles/Diagnostics` - `AutoForge.UI/Features/Vehicles/Simulation` No mediator, generic repository layer, database, or empty future feature is required. See the [schema-v1 profile guide](docs/vehicle-profile-schema-v1.md) for stable identifiers/revisions, unknown values, capabilities, provenance and the strict JSON contract. ## Repository layout Only `docs`, `src` and `prompts` are top-level project directories. Tests and synthetic fixtures remain under `src/tests`. No nested repository or root test project is used. ```text AutoForge/ docs/ Technical docs, schemas and ADRs prompts/ Manual assignments, state and worker reports src/ AutoForge.slnx AutoForge.Client/ Sole Windows MAUI Hybrid executable AutoForge.UI/ Application routes and feature presentation AutoForge.Components/ Generic reusable UI components AutoForge.Core/ Platform-neutral domain invariants AutoForge.Application/ In-process application feature slices AutoForge.Diagnostics/ Provider-neutral diagnostic protocol/contracts AutoForge.Simulation/ Static modules, dynamics engine and transports tests/ AutoForge.UnitTests/ AutoForge.IntegrationTests/ Fixtures/ Clearly synthetic schema-v1 JSON only ``` ## Build, test and run Verified Windows prerequisites: .NET SDK **10.0.402**, runtime **10.0.13**, Windows SDK target **10.0.26100.0**, `maui-windows` workload manifest **10.0.110** and WebView2 Runtime **155.0.4283.62**. The Client targets `net10.0-windows10.0.26100.0`/`win-x64`; the other projects target portable `net10.0`. No automotive driver, actual vehicle, AI model, database, backend or cloud account is needed. From the repository root in PowerShell: ```powershell dotnet restore .\src\AutoForge.slnx dotnet build .\src\AutoForge.slnx --configuration Release --no-restore dotnet test .\src\AutoForge.slnx --configuration Release --no-build dotnet run --project .\src\AutoForge.Client\AutoForge.Client.csproj --configuration Release --framework net10.0-windows10.0.26100.0 ``` Step-02A worker verification includes a seven-project Release build with zero warnings/errors, 103 unit and 37 integration tests passed (140 total, zero failed/skipped), and native Client checks for strict timestamp failures, partial evidence import/export and conflicting-identity atomicity. The original Step-02 report records its historical 59 unit/34 integration cases; those runs did not include the Step-02A corrections. Step-03 extends those suites and records its actual full-build/test and native-smoke evidence in the [Step-03 worker report](prompts/Step-03-Report.md). Step-03A's stricter request/reply validation, trace consistency and clear-verification behavior, tests, final 111-unit/64-integration Release results, and worker evidence are recorded in the [Step-03A corrective report](prompts/Step-03A-Report.md). Manual Step-03 acceptance remains the orchestrator's responsibility. ## Documentation and handoff - [Documentation index](docs/README.md) - [Architecture and project graph](docs/architecture.md) - [Synthetic diagnostics protocol v1](docs/synthetic-diagnostics-v1.md) - [Synthetic diagnostics protocol v2](docs/synthetic-diagnostics-v2.md) - [Simulation definitions and state](docs/simulation.md) - [Live dashboard usage and contracts](docs/live-dashboard.md) - [Telemetry archive schema v1 and recording](docs/telemetry-archive-v1.md) - [Recorded telemetry replay](docs/telemetry-replay.md) - [Synthetic calibration authoring](docs/calibration-authoring.md) - [Verified simulated programming rehearsal](docs/calibration-rehearsal.md) - [Manual runbook schema v1 and read-only preview](docs/runbooks-v1.md) - [Local-only model generation and prerequisites](docs/local-runbook-generation.md) - [Vehicle-profile schema v1 and usage](docs/vehicle-profile-schema-v1.md) - [Build, tests and native smoke guide](docs/build-and-test.md) - [ADR-0001: historical .NET foundation](docs/adr-0001-headless-foundation.md) - [ADR-0002: one native shell and feature slices](docs/adr-0002-single-shell-vsa.md) - [ADR-0003: strict profile interchange v1](docs/adr-0003-profile-interchange-v1.md) - [ADR-0004: synthetic diagnostics boundaries](docs/adr-0004-synthetic-diagnostics.md) - [Step-01 historical report](prompts/Step-01-Report.md) - [Accepted Step-01A report](prompts/Step-01A-Report.md) - [Step-02 worker report](prompts/Step-02-Report.md) - [Step-02A corrective worker report](prompts/Step-02A-Report.md) - [Step-03 worker report](prompts/Step-03-Report.md) - [Step-03A corrective worker report](prompts/Step-03A-Report.md) - [Step-03B corrective worker report](prompts/Step-03B-Report.md) - [Step-04 worker report](prompts/Step-04-Report.md) - [Step-04A corrective worker report](prompts/Step-04A-Report.md) - [Step-04B corrective worker report](prompts/Step-04B-Report.md) - [Step-07 worker report](prompts/Step-07-Report.md) - [Step-05A corrective worker report](prompts/Step-05A-Report.md) - [Orchestrator-owned roadmap](prompts/ROADMAP.md) and [orchestration state](prompts/STATE.md) Handoffs are manual. Worker completion does not mean orchestrator acceptance or authorize subsequent work.