CodeHear is a single VS Code extension for blind, visually impaired, and low-vision developers.
It narrates code structure (functions, classes, loops, call paths, diagnostics, and scope) so you can navigate code semantically instead of only hearing raw punctuation and line text.
What You Get (User POV)
- Faster orientation in unknown files.
- Deterministic structural narration while moving through code.
- Dedicated ecosystem modes for repo understanding, debugging, review, testing, docs, and focus workflows.
- Guided missions that announce "Step X of Y" progress.
- Local-only AI reasoning via Ollama for optional, explicit commands.
- Bookmarks, landmarks, macros, and accessibility profile export/import.
Core Guarantees
- Local-first deterministic parser-driven narration.
- Optional AI is local Ollama only (localhost), never cloud by default.
- Real-time cursor narration does not depend on AI.
- Single interruptible speech channel with dedupe/throttling.
- Graceful fallback speech when syntax is incomplete or parsers fail.
Supported Languages
- JavaScript (
javascript,javascriptreact) - TypeScript (
typescript,typescriptreact) - Python (
python) - Java (
java) - C (
c) - C++ (
cpp) - Go (
go) - Rust (
rust)
Install, Build, Validate
npm install
npm run compile
npm test
Optional full sanity workflow:
npm run sanity:first30
First Run (5 Minutes)
- Press
F5to launch Extension Development Host. - Open a supported code file.
- Move the cursor and run
CodeHear: Describe Current Context. - Run
CodeHear: Quick Context. - Run
CodeHear: File Orientation Summary. - Switch modes with
CodeHear: Switch Ecosystem Mode.
Voice Commands (Complete)
Important: CodeHear currently supports voice intent through the CodeHear: Switch Mode By Voice Intent command.
- It opens an input box.
- You can speak via OS dictation or type the phrase.
- CodeHear parses your phrase into one ecosystem mode.
hearandherevariants are accepted for all Hear modes (for examplerepo hearandrepo here).- If voice input is not captured or cannot be mapped, CodeHear opens an accessible mode picker automatically.
Accepted Voice Intent Phrases By Mode
| Mode | Accepted keywords and phrases |
|---|---|
| IDE | ide, editor, code mode, coding |
| RepoHear | repohear, repo hear, repo, repository |
| DebugHear | debughear, debug hear, debug, debugging |
| ReviewHear | reviewhear, review hear, review, code review, pr review |
| TestHear | testhear, test hear, test, testing |
| LearnHear | learnhear, learn hear, learn, learning |
| TerminalHear | terminalhear, terminal hear, terminal, console |
| GitHear | githear, git hear, git, source control |
| DocHear | dochear, doc hear, docs, documentation |
| SecurityHear | securityhear, security hear, security, secure |
| PerfHear | perfhear, perf hear, performance, perf |
| CollabHear | collabhear, collab hear, collab, collaboration, handoff |
| FocusHear | focushear, focus hear, focus, low vision |
Voice Intent Examples
- "Switch to DebugHear mode"
- "Go to repo mode"
- "Switch to repo here mode"
- "Enable FocusHear"
- "Switch to PR review"
Ecosystem Modes And How To Use Them
| Mode | Best for | Start with |
|---|---|---|
| IDE | Core structural coding navigation | codeHear.quickContext, codeHear.moveParent |
| RepoHear | Understanding architecture and entrypoints | codeHear.understandRepo, codeHear.repoOverviewLocal |
| DebugHear | Diagnostics and bug triage | codeHear.jumpToNextDiagnostic, codeHear.startDebugCopilotLoop |
| ReviewHear | Risk-first changed file review | codeHear.reviewChangedFiles, codeHear.reviewRiskSummaryLocal |
| TestHear | Interpreting failing tests | codeHear.testFailureExplanationLocal, codeHear.jumpToNextDiagnostic |
| LearnHear | Beginner-friendly contextual learning | codeHear.quickContext, codeHear.summarizeIntentLocal |
| TerminalHear | Translating terminal output into actionable steps | codeHear.testFailureExplanationLocal, codeHear.quickContext |
| GitHear | Commit readiness and handoff | codeHear.reviewChangedFiles, codeHear.generateHandoffSummary |
| DocHear | Setup/troubleshooting/document navigation | codeHear.repoOverviewLocal, codeHear.intentQuery |
| SecurityHear | Security smell prioritization | codeHear.reviewRiskSummaryLocal, codeHear.explainCurrentDiagnostic |
| PerfHear | Performance hotspot awareness | codeHear.audioRadar, codeHear.summarizeFunctionChangesSinceSave |
| CollabHear | Team sync and handoff output | codeHear.generateHandoffSummary, codeHear.missionStatus |
| FocusHear | Reduced-clutter low-vision coding | codeHear.quickContext, codeHear.fileOrientationSummary |
Complete Command Catalog (All Features)
1) Structural Navigation And Context
| Command ID | Command Palette Title | What it does | How to use |
|---|---|---|---|
codeHear.describeCurrentContext |
CodeHear: Describe Current Context | Speaks the semantic unit at cursor. | Run on any code line to hear current block meaning. |
codeHear.quickContext |
CodeHear: Quick Context | Speaks compact mode + task + location context. | Use as a frequent orientation checkpoint. |
codeHear.moveParent |
CodeHear: Move To Parent Node | Moves to parent AST node and narrates. | Run when you want to go one scope up. |
codeHear.moveFirstChild |
CodeHear: Move To First Child Node | Moves to first child node and narrates. | Run to drill into current block quickly. |
codeHear.moveNextSibling |
CodeHear: Move To Next Sibling | Moves to next sibling node. | Run for lateral navigation in same scope. |
codeHear.movePreviousSibling |
CodeHear: Move To Previous Sibling | Moves to previous sibling node. | Run to backtrack sibling-level traversal. |
codeHear.selectCurrentBlock |
CodeHear: Select Current Structural Block | Selects full current semantic block. | Run before refactor/copy actions. |
codeHear.jumpNextFunction |
CodeHear: Jump To Next Function | Jumps to next function or method. | Use to skim implementation order. |
codeHear.jumpNextClass |
CodeHear: Jump To Next Class | Jumps to next class block. | Use in OO-heavy files. |
codeHear.jumpNextLoop |
CodeHear: Jump To Next Loop | Jumps to next loop construct. | Use for performance and flow review. |
codeHear.jumpNextErrorNode |
CodeHear: Jump To Next Error Node | Jumps to parse error region. | Use when file has incomplete syntax. |
codeHear.fileOrientationSummary |
CodeHear: File Orientation Summary | Speaks classes/functions/loops/diagnostics map. | Run when opening unfamiliar files. |
codeHear.audioRadar |
CodeHear: Audio Radar | Plays optional earcon radar then structural summary. | Run for compact high-level topology. |
codeHear.speakBreadcrumbPath |
CodeHear: Speak Breadcrumb Path | Speaks AST breadcrumb chain. | Use to understand nested ownership. |
codeHear.speakCallPath |
CodeHear: Speak Call Path | Speaks call-path around selected symbol. | Use during bug tracing and design review. |
codeHear.intentQuery |
CodeHear: Intent Query | Finds likely code locations for natural-language query. | Ask "where is token validated" style questions. |
codeHear.summarizeFunctionChangesSinceSave |
CodeHear: Summarize Function Changes Since Save | Reports unsaved changes and risk signals in current scope. | Run before save/commit. |
codeHear.repeatLastAnnouncement |
CodeHear: Repeat Last Announcement | Repeats previous spoken output. | Use when speech was interrupted or missed. |
codeHear.spellCurrentToken |
CodeHear: Spell Current Token | Spells current token with optional phonetic map. | Use for exact symbol verification. |
codeHear.enterStickyCommandMode |
CodeHear: Enter Sticky Command Mode | Keeps command picker loop active for rapid nav. | Enter for repeated command bursts. |
codeHear.exitStickyCommandMode |
CodeHear: Exit Sticky Command Mode | Exits sticky command loop. | Run when finished burst actions. |
codeHear.repeatLastAction |
CodeHear: Repeat Last Action | Re-runs last command action. | Use for repetitive movement/check loops. |
2) Diagnostics And Debug
| Command ID | Command Palette Title | What it does | How to use |
|---|---|---|---|
codeHear.jumpToNextDiagnostic |
CodeHear: Jump To Next Diagnostic | Jumps cursor to next diagnostic. | Run to triage compile/lint errors quickly. |
codeHear.explainCurrentDiagnostic |
CodeHear: Explain Current Diagnostic | Speaks likely cause and fix pattern. | Run after jumping to an error. |
codeHear.startDebugCopilotLoop |
CodeHear: Start Debug Copilot Loop | Runs narrated debug sequence end-to-end. | Use when actively fixing a bug. |
codeHear.explainCurrentStackFrame |
CodeHear: Explain Current Stack Frame | Narrates current stack frame context. | Run in debug investigations. |
codeHear.summarizeVariableChanges |
CodeHear: Summarize Variable Changes | Narrates variable deltas around cursor context. | Use while stepping or reviewing state. |
codeHear.suggestNextBreakpoint |
CodeHear: Suggest Next Breakpoint | Suggests useful next breakpoint location. | Run after first failure localization. |
codeHear.speakCallStackBreadcrumb |
CodeHear: Speak Call Stack Breadcrumb | Speaks call-stack breadcrumb string. | Use for stack orientation mid-debug. |
3) Ecosystem Modes
| Command ID | Command Palette Title | What it does | How to use |
|---|---|---|---|
codeHear.switchMode |
CodeHear: Switch Ecosystem Mode | Opens picker and activates selected mode. | Main mode switch method. |
codeHear.switchModeByVoice |
CodeHear: Switch Mode By Voice Intent | Parses spoken/typed phrase into mode. | Use phrase like "switch to debughear". |
codeHear.currentMode |
CodeHear: Speak Current Mode | Speaks active mode and quick context. | Run when unsure of active context. |
codeHear.openActionCenter |
CodeHear: Open Mini Action Center | Opens compact action hub to run common extension commands. | Use as one-place command launcher. |
codeHear.listModes |
CodeHear: List Ecosystem Modes | Opens mode output panel with descriptions and actions. | Run to learn mode purpose and entry cues. |
4) Mission Engine
| Command ID | Command Palette Title | What it does | How to use |
|---|---|---|---|
codeHear.startMission |
CodeHear: Start Mission | Starts Bug-Fix, Feature, Review, or Test-Fix mission template. | Select template, then enter mission goal. |
codeHear.nextMissionStep |
CodeHear: Next Mission Step | Advances mission to next step. | Run after completing current step. |
codeHear.repeatMissionStep |
CodeHear: Repeat Mission Step | Repeats current mission instruction. | Run when you need current step repeated. |
codeHear.missionStatus |
CodeHear: Mission Status | Speaks template, step, and completion state. | Use in long sessions to re-sync. |
codeHear.completeMission |
CodeHear: Complete Mission | Marks mission complete and narrates completion. | Run when mission goal is done. |
5) Local AI (Ollama, Localhost Only)
| Command ID | Command Palette Title | What it does | How to use |
|---|---|---|---|
codeHear.setupLocalAiGuided |
CodeHear: Setup Local AI Guided | Guided setup for local Ollama endpoint/model. | Run once during initial setup. |
codeHear.checkLocalAiStatus |
CodeHear: Check Local AI Status | Reports health, model, and installed tags. | Run when local AI commands fail. |
codeHear.selectLocalAiModel |
CodeHear: Select Local AI Model | Picks installed model and updates settings. | Run after pulling new model. |
codeHear.summarizeIntentLocal |
CodeHear: Summarize Intent (Local Ollama) | Gives intent summary for current semantic node. | Run when learning unfamiliar blocks. |
codeHear.pullLocalAiModel |
CodeHear: Pull Local AI Model | Starts ollama pull in integrated terminal and syncs model settings. |
Run when adding new model locally. |
codeHear.repoOverviewLocal |
CodeHear: Repo Overview (Local Reasoning) | Generates repository orientation narration. | Use for onboarding in unknown repo. |
codeHear.reviewRiskSummaryLocal |
CodeHear: Review Risk Summary (Local Reasoning) | Summarizes PR risk in concise spoken form. | Run before review or merge. |
codeHear.testFailureExplanationLocal |
CodeHear: Test Failure Explanation (Local Reasoning) | Explains pasted test failure text. | Paste assertion/trace and run command. |
codeHear.generateHandoffSummary |
CodeHear: Generate Accessibility Handoff Summary | Creates spoken + markdown team handoff artifact. | Run at end of session or shift handoff. |
6) Task, Repo, Review, Productivity
| Command ID | Command Palette Title | What it does | How to use |
|---|---|---|---|
codeHear.setTaskMode |
CodeHear: Set Task Mode | Sets read/debug/refactor/pr-review mode and narration presets. | Run at start of a focused task. |
codeHear.startTask |
CodeHear: Start Task | Generates and narrates multi-step task flow from your goal. | Enter goal like "find why login fails". |
codeHear.understandRepo |
CodeHear: Understand This Repo | Produces architecture/dataflow/hotspot/entrypoint summary. | Use during repo onboarding. |
codeHear.reviewChangedFiles |
CodeHear: Review Changed Files | Risk-ranks changed files and narrates review order. | Run before code review/commit. |
codeHear.showProductivityMetrics |
CodeHear: Show Productivity Metrics | Shows session productivity metrics panel. | Run to track workflow gains. |
codeHear.showRecentLandmarks |
CodeHear: Show Recent Landmarks | Lists recent visited semantic landmarks. | Use to jump back to recent context. |
codeHear.addBookmarkWithLabel |
CodeHear: Add Bookmark With Label | Saves labeled bookmark at current position. | Add stable navigation anchors. |
codeHear.jumpToBookmark |
CodeHear: Jump To Bookmark | Jumps to selected saved bookmark. | Use for fast context switching. |
codeHear.runMacro |
CodeHear: Run Macro | Executes pre-defined macro command sequence. | Use for repeatable command routines. |
codeHear.selectVoicePackTheme |
CodeHear: Select Voice Pack And Earcon Theme | Updates voice pack and earcon theme presets. | Use to tune fatigue and audio style. |
7) Accessibility Profile Management
| Command ID | Command Palette Title | What it does | How to use |
|---|---|---|---|
codeHear.exportProfile |
CodeHear: Export Accessibility Profile | Exports accessibility settings as JSON. | Use to back up/share setup. |
codeHear.importProfile |
CodeHear: Import Accessibility Profile | Imports profile JSON to workspace settings. | Use to apply team/user presets quickly. |
8) Dynamic Intelligence Layer (A-I)
| Command ID | Command Palette Title | What it does | How to use |
|---|---|---|---|
codeHear.currentIntentState |
CodeHear: Current Intent State | Speaks current inferred intent state and confidence. | Run to verify what CodeHear thinks you are doing. |
codeHear.explainWhyCurrentState |
CodeHear: Explain Why Current Intent State | Speaks intent state plus reasons from recent signals. | Run when state feels wrong and you want reasoning. |
codeHear.suggestNextAction |
CodeHear: Suggest Next Action | Speaks top next-best action from dynamic queue. | Run as lightweight guidance during debug/review loops. |
codeHear.acceptSuggestedAction |
CodeHear: Accept Suggested Action | Executes the top queued action (or speaks fallback). | Run to confirm queue recommendation immediately. |
codeHear.dismissSuggestedAction |
CodeHear: Dismiss Suggested Action | Removes top queued action. | Run when suggestion is low-value for current goal. |
codeHear.readSuggestionQueue |
CodeHear: Read Suggestion Queue | Opens output queue and narrates top suggestion. | Run to inspect ranked queue before acting. |
codeHear.currentRiskPulse |
CodeHear: Current Risk Pulse | Speaks current risk level, score, and confidence. | Run before commits or when edits feel risky. |
codeHear.explainRiskPulse |
CodeHear: Explain Risk Pulse | Explains why pulse is high/medium/low. | Run to understand risk drivers in current file. |
codeHear.jumpToTopRiskRegion |
CodeHear: Jump To Top Risk Region | Moves cursor to highest-risk tracked region. | Run for risk-first review and fix order. |
codeHear.enableMissionAutopilot |
CodeHear: Enable Mission Autopilot | Turns on automatic mission replanning. | Enable for long multi-step debugging sessions. |
codeHear.disableMissionAutopilot |
CodeHear: Disable Mission Autopilot | Turns off automatic mission replanning. | Disable for strict manual mission control. |
codeHear.contextRescueNow |
CodeHear: Context Rescue Now | Speaks drift recovery summary with landmarks and intent. | Run when feeling disoriented in deep code. |
codeHear.toggleAutoContextRescue |
CodeHear: Toggle Auto Context Rescue | Toggles automatic rescue trigger. | Run to control proactive rescue behavior. |
codeHear.predictNextLocations |
CodeHear: Predict Next Locations | Generates top predicted next jump targets. | Run before repeated diagnostic/changed-line jumps. |
codeHear.jumpPredictedLocation1 |
CodeHear: Jump Predicted Location 1 | Jumps to top prediction if confidence passes threshold. | Use with keyboard-only prediction workflow. |
codeHear.jumpPredictedLocation2 |
CodeHear: Jump Predicted Location 2 | Jumps to second prediction if confidence passes threshold. | Use when first prediction is not relevant. |
codeHear.jumpPredictedLocation3 |
CodeHear: Jump Predicted Location 3 | Jumps to third prediction if confidence passes threshold. | Use for broader predicted triage sweep. |
codeHear.handoffReadiness |
CodeHear: Handoff Readiness | Scores readiness and narrates blockers/actions. | Run before shift handoff or PR transfer. |
Keybindings (Defaults)
CodeHear ships preset-based keybindings:
codeHear.keybindingPreset = nvda-friendlycodeHear.keybindingPreset = voiceover-friendlycodeHear.keybindingPreset = custom
High-use defaults (nvda-friendly) include:
Ctrl+Alt+D: Describe Current ContextCtrl+Alt+Q: Quick ContextCtrl+Alt+S: Switch ModeCtrl+Alt+W: Switch Mode By Voice IntentCtrl+Alt+0: Open Mini Action CenterCtrl+Alt+Y: Speak Current ModeCtrl+Alt+E: Jump To Next DiagnosticCtrl+Alt+M: Enter Sticky Command ModeCtrl+Alt+R: Repeat Last Announcement
Workflow-gated defaults:
Ctrl+Alt+T: Start Task (read mode)Ctrl+Alt+G: Start Debug Copilot Loop (debug mode)Ctrl+Alt+V: Review Changed Files (pr-review mode)
Mission defaults:
Ctrl+Alt+Shift+T: Start MissionCtrl+Alt+]: Next Mission StepCtrl+Alt+[: Mission Status
Dynamic defaults:
Ctrl+Alt+1: Current Intent StateCtrl+Alt+2: Suggest Next ActionCtrl+Alt+3: Accept Suggested ActionCtrl+Alt+4: Dismiss Suggested ActionCtrl+Alt+5: Current Risk PulseCtrl+Alt+6: Jump To Top Risk RegionCtrl+Alt+7: Context Rescue NowCtrl+Alt+8: Predict Next LocationsCtrl+Alt+9: Handoff ReadinessCtrl+Alt+Shift+1: Jump Predicted Location 1Ctrl+Alt+Shift+2: Jump Predicted Location 2Ctrl+Alt+Shift+3: Jump Predicted Location 3
How To Use Every Feature Group (Practical Flows)
Flow A: Daily Coding Orientation
- Run
codeHear.switchModeand chooseideorfocushear. - Run
codeHear.quickContextevery time you enter a new file/section. - Use
codeHear.moveParent,codeHear.moveNextSibling, andcodeHear.moveFirstChildfor structural movement. - Use
codeHear.fileOrientationSummarybefore larger edits.
Flow B: Fast Bug Fix
- Run
codeHear.switchModeByVoicewith "switch to debughear". - Run
codeHear.startMissionwith templatebug-fix. - Run
codeHear.startDebugCopilotLoop. - Run
codeHear.nextMissionStepas each checkpoint is finished. - Run
codeHear.completeMissionwhen fixed.
Flow C: PR Review
- Run
codeHear.switchModeand selectreviewhear. - Run
codeHear.reviewChangedFiles. - Run
codeHear.reviewRiskSummaryLocal(if local AI enabled). - Run
codeHear.generateHandoffSummaryfor reviewer handoff notes.
Flow D: Test Failure Triage
- Run
codeHear.switchModeand selecttesthearorterminalhear. - Run
codeHear.testFailureExplanationLocaland paste failing output. - Run
codeHear.jumpToNextDiagnosticandcodeHear.explainCurrentDiagnostic. - Run
codeHear.summarizeFunctionChangesSinceSavebefore re-run.
Flow E: Repo Onboarding
- Run
codeHear.switchModeByVoicewith "switch to repohear". - Run
codeHear.understandRepo. - Run
codeHear.repoOverviewLocal. - Run
codeHear.intentQueryfor targeted search (for example "where is auth token validated").
Flow F: Dynamic Rescue and Risk-First Fixing
- Run
codeHear.currentIntentStateto confirm inferred task state. - Run
codeHear.suggestNextActionandcodeHear.acceptSuggestedActionfor guided next step. - Run
codeHear.currentRiskPulseandcodeHear.jumpToTopRiskRegionto prioritize highest-risk edits. - Run
codeHear.predictNextLocationsthen jump viacodeHear.jumpPredictedLocation1/2/3. - Run
codeHear.contextRescueNowwhen you lose track. - Run
codeHear.handoffReadinessbefore handing work to teammate/reviewer.
Settings You Will Use Most
Narration and Accessibility
codeHear.verbositycodeHear.verbosityProfilecodeHear.punctuationModecodeHear.pushToTalkcodeHear.coexistenceModecodeHear.announcementDedupcodeHear.announcementThrottleMs
Modes and Workflow
codeHear.mode.defaultcodeHear.mode.persistPerWorkspacecodeHear.taskModecodeHear.taskModeAutoPreset
Voice and Earcons
codeHear.voice.ratecodeHear.voice.pitchcodeHear.voice.volumecodeHear.voice.preferredNamecodeHear.voice.pausePunctuationcodeHear.voice.stylecodeHear.voicePackcodeHear.earconTheme
Local AI (Optional)
codeHear.localAI.enabledcodeHear.localAI.providercodeHear.localAI.baseUrlcodeHear.localAI.modelcodeHear.localAI.timeoutMscodeHear.localAI.maxTokenscodeHear.localAI.temperaturecodeHear.localAI.spokenNoticeEnabled
Dynamic Intelligence
codeHear.dynamicNarration.enabledcodeHear.dynamicNarration.maxShiftPerMinutecodeHear.fatigueGuard.enabledcodeHear.fatigueGuard.checkpointMinutescodeHear.intentEngine.enabledcodeHear.intentEngine.transitionDebounceMscodeHear.suggestionEngine.enabledcodeHear.suggestionEngine.maxQueueSizecodeHear.riskPulse.enabledcodeHear.riskPulse.announceThresholdcodeHear.riskPulse.highRiskInterruptEnabledcodeHear.missionAutopilot.enabledcodeHear.contextRescue.autoEnabledcodeHear.contextRescue.cooldownMscodeHear.sonar.enabledcodeHear.sonar.intensitycodeHear.sonar.depthMappingcodeHear.sonar.errorProximityAlertscodeHear.prediction.enabledcodeHear.prediction.minConfidencecodeHear.handoffReadiness.enabled
Dynamic Triggers and Benefits
The dynamic layer continuously watches navigation, command usage, diagnostics, changed lines, mission state, and context drift signals.
- Blind developers: Faster recovery from disorientation loops via proactive context rescue and intent narration.
- Visually impaired developers: Lower cognitive load through confidence-ranked next actions and risk-first jumps.
- Low-vision developers: Better sustained focus with fatigue checkpoints, adaptive verbosity, and sonar cues.
High-value trigger examples:
- Repeated diagnostic jumps plus undo bursts: intent shifts toward debug.
- Rapid file-switch loops with backtracking: context rescue trigger candidates.
- Security-sensitive changed lines plus rising diagnostics: risk pulse threshold interrupt.
- Mission in progress with scope/diagnostic/test changes: autopilot mission replan.
Deterministic Fallback Behavior
- Local AI failures never block command completion.
- Suggestions, risk explanations, and test-failure explanations fall back to deterministic local templates.
- If Ollama is unavailable, CodeHear narrates actionable fix guidance and continues with non-AI output.
- Push-to-talk and strict screen-reader coexistence still suppress auto narration paths exactly as before.
Local AI Policy
- Local AI is optional.
- Local AI endpoints must be localhost (
localhost,127.0.0.1,::1). - Remote endpoints are blocked by design.
- Cursor-move narration remains deterministic and non-AI.
Additional Guides
USER_POV_EXTENSION_GUIDE.mdACCESSIBILITY_GUIDE.mdMODES_GUIDE.mdMISSIONS_GUIDE.mdOLLAMA_SETUP.mdTROUBLESHOOTING.md
Architecture Entry Points
- Activation and runtime wiring:
src/extension.ts - Structural command layer:
src/commands/structuralCommands.ts - Workflow command layer:
src/commands/workflowCommands.ts - Ecosystem/mode/mission/local-AI command layer:
src/commands/ecosystemCommands.ts - Dynamic intent/risk/rescue/prediction command layer:
src/commands/dynamicCommands.ts - Mode engine:
src/workflows/modeManagerService.ts - Mission engine:
src/workflows/missionEngineService.ts - Intent state engine:
src/workflows/intentStateEngine.ts - Suggestion engine:
src/workflows/suggestionEngine.ts - Risk pulse service:
src/workflows/riskPulseService.ts - Context rescue service:
src/workflows/contextRescueService.ts - Predictive jump service:
src/workflows/predictiveJumpService.ts - Handoff readiness service:
src/workflows/handoffReadinessService.ts - Sonar cue service:
src/audio/sonarService.ts - Local AI service:
src/ai/localAiService.ts
Current Validation Snapshot
Latest validation run in this repository:
npm run compile: passingnpm run lint: passingnpm test: passing (53 passing,0 failing)npm run sanity:first30: completed all scripted checks successfully