feat(languages): add Rust, C++, and Go across all 452 algorithms

.claude/CLAUDE.md

@@ -24,21 +24,14 @@ Registry-driven, pre-computed steps, tracker abstraction, discriminated `VisualS
 - Types: `src/types/`
 - Registry: `src/registry/`
 - Trackers: `src/trackers/`
-- Algorithms: `src/algorithms/<category>/<algorithm>/`
+- Algorithms: `src/algorithms/<category>/<technique>/<algorithm>/`
 - Store: `src/store/`
 - Components: `src/components/`
 - Plan: `.claude/PLAN.md`
 
 ## Rules
 
-See `.claude/rules/` for full constraints. Most commonly violated:
-
-- No single-char variable names, no `any` types — use `unknown` with narrowing
-- `@/` path alias required for all src-relative imports
-- `noUncheckedIndexedAccess` enabled — use tuple types (`[number, number][]`) for coordinate arrays
-- Pipeline stories (`*.Pipeline.stories.tsx`) live in algorithm directories, not `src/components/`
-- Branch-per-task mandatory — every new task starts on a fresh branch from main
-- All edits to input and pathfinding grids are temporary (non-persistent)
+See `.claude/rules/` for full constraints. Most commonly violated: naming (no single-char), no `any`, `@/` imports, `noUncheckedIndexedAccess` tuple types, `__tests__/` for stories, branch-per-task, non-persistent edits.
 
 ## General Guidelines
 
@@ -53,14 +46,11 @@ See `.claude/rules/` for full constraints. Most commonly violated:
 
 ## Git Workflow
 
-- When working on branches, always verify the current branch and its commit history before starting work. Never create a new branch from main when continuing prior work — check for existing feature branches first.
+Branch-per-task mandatory. Check existing feature branches before creating new ones. See `.claude/rules/workflow.md`.
 
 ## Testing
 
-- Run CI checks in sequence and fix iteratively until all pass green: `npm run lint` → `npm run format` → `npm run typecheck` → `npm test`
-- Do not stop after fixing just one category — keep going until everything is clean
-- For E2E tests, the dev server starts automatically via hooks — do not start it manually
-- Every new algorithm needs: correctness tests + step generation tests + pipeline story in algorithm directory + E2E registration — see `.claude/rules/testing.md`
+CI sequence: `lint` → `format` → `typecheck` → `test` — fix iteratively until all green. Dev server auto-starts for E2E. Every new algorithm needs tests + story in `__tests__/`. See `.claude/rules/testing.md`.
 
 ## Workflow
 

.claude/PLAN.md

@@ -4,7 +4,7 @@
 
 Build a learner-focused algorithm visualization web app from scratch. The app provides synchronized code-line highlighting with step-by-step algorithm execution, multi-language code display, interactive pathfinding grid editing, and rich educational content. The project includes a full `.claude` system (rules, agents, skills, hooks) for maintainable development workflow.
 
-**Working directory**: `/Users/springfield/dev/algo_flow` (currently empty, not a git repo)
+**Working directory**: `/Users/springfield/dev/algo_flow`
 
 ---
 
@@ -89,7 +89,6 @@ algo_flow/
 │   │   ├── tech-lead-architect.md
 │   │   ├── product-strategist.md
 │   │   ├── technical-writer.md
-│   │   ├── marketing-engine.md
 │   │   ├── claude-system-architect.md
 │   │   ├── silent-failure-hunter.md
 │   │   ├── code-simplifier.md
@@ -98,9 +97,7 @@ algo_flow/
 │   │   ├── implementation-planning/SKILL.md
 │   │   ├── algorithm-learning-content/SKILL.md
 │   │   ├── pathfinding-scenario-editing/SKILL.md
-│   │   ├── repository-quality-gate/SKILL.md
 │   │   ├── branch-safety-check/SKILL.md
-│   │   ├── cifix/SKILL.md
 │   │   ├── accessibility-audit/SKILL.md
 │   │   ├── architecture-review/SKILL.md
 │   │   ├── strict-typescript-review/SKILL.md
@@ -115,23 +112,30 @@ algo_flow/
 │   │   └── debugging/SKILL.md
 │   └── hooks/
 │       ├── session-start-branch-check.sh
+│       ├── session-end-unified-gate.sh
 │       ├── session-end-quality-gate.sh
 │       ├── session-end-readme-check.sh
 │       ├── session-end-comments-check.sh
 │       ├── session-end-e2e-check.sh
 │       ├── session-end-security-check.sh
 │       ├── session-end-claude-system-check.sh
+│       ├── auto-plugin-mode.sh
+│       ├── auto-pr-after-push.sh
+│       ├── ban-hardcoded-waits.sh
 │       ├── block-ai-attribution.sh
 │       ├── block-main-branch-commits.sh
+│       ├── enforce-branch-naming.sh
+│       ├── pre-commit-fn-import-check.sh
 │       ├── pre-commit-quality-check.sh
-│       ├── auto-pr-after-push.sh
-│       ├── post-edit-typescript-check.sh
-│       └── post-edit-accessibility-check.sh
+│       ├── post-edit-accessibility-check.sh
+│       ├── post-edit-java-check.sh
+│       ├── post-edit-python-check.sh
+│       └── post-edit-typescript-check.sh
 ├── src/
 │   ├── types/                       # All TypeScript interfaces
 │   ├── registry/                    # AlgorithmRegistry singleton
 │   ├── engine/                      # Step generator
-│   ├── trackers/                    # BaseTracker + 6 category trackers
+│   ├── trackers/                    # BaseTracker + category subdirectories
 │   ├── store/                       # Zustand 4-slice store
 │   ├── algorithms/
 │   │   ├── sorting/bubble-sort/     # definition, algo, steps, educational, sources/
@@ -199,7 +203,7 @@ algo_flow/
 ### Phase 4: Bubble Sort (Full Pipeline Proof-of-Concept)
 
 - Pure `bubbleSort()` implementation + unit tests
-- Source files: TypeScript, Python, Java
+- Source files: TypeScript, Python, Java, Rust, C++, Go
 - Step generator using SortingTracker + line map
 - Educational content (all 7 sections)
 - CodePanel with Monaco: read-only, language tabs, line highlighting
@@ -277,6 +281,3 @@ algo_flow/
 
 ---
 
-## File Count Estimate
-
-~206 files total: 22 `.claude/` config, 8 types, 4 registry/engine, 12 trackers, 8 store, 48 algorithm files, 18 source display files, 50 components+stories, 8 hooks, 5 utils, 5 e2e, 12 config, 6 infra.

.claude/agents/qa-tester.md

@@ -20,7 +20,7 @@ Validate that all features work correctly and test coverage meets thresholds.
 4. **Playback**: Play, pause, step, speed, reset, rerun all function
 5. **Input editing**: Temporary edits trigger recompute, reset on algorithm switch
 6. **Pathfinding editing**: Wall toggle, start/end drag, run, reset all work. Edits non-persistent.
-7. **Language switching**: Code panel updates correctly for all 3 languages
+7. **Language switching**: Code panel updates correctly for all 6 languages
 8. **Responsive layout**: Works at desktop, tablet, mobile breakpoints
 
 ## Test Execution
@@ -29,20 +29,20 @@ Validate that all features work correctly and test coverage meets thresholds.
 - Run `npm run lint` and report results
 - Run `npm run format:check` and report results
 - Run `npm run typecheck` and report results
-- Verify coverage meets thresholds (80/75/80/80)
+- Verify coverage meets thresholds per `rules/testing.md` (80/75/80/80)
 
 ## Required Skills
 
 - **Playwright E2E**: Multi-viewport testing (1280/768/375), algorithm flows, keyboard shortcuts
-- **Coverage enforcement**: 80/75/80/80 thresholds
-- **OWASP client-side**: XSS prevention, dependency audit — see `security-coverage-audit` skill for detailed checklist
+- **Coverage enforcement**: Per `rules/testing.md`
+- **OWASP client-side**: XSS prevention, dependency audit — see `security-coverage-audit` skill
 
 ## Constraints
 
 - Never approve a PR with coverage below thresholds without explicit justification
 - E2E tests must cover all 3 viewports for any new visual component
 - Security checks must include `npm audit` and manual review of any new dynamic content rendering
-- All algorithm additions must be importable via `src/algorithms/index.ts` so per-category E2E spec files auto-discover them; algorithms with custom input editors must have an entry in `e2e/specs/input-editors.spec.ts`
+- E2E auto-discovers from registry — see `rules/testing.md` for spec file convention
 
 ## Output Format
 

.claude/agents/senior-engineer-code-reviewer.md

@@ -19,8 +19,8 @@ Review code changes for quality, correctness, and adherence to project standards
 3. **Types**: Proper TypeScript usage. No `any`. Discriminated unions used correctly.
 4. **DRY**: No duplicated logic. Reused strings centralized in constants.
 5. **Tests**: Algorithm implementations have unit tests. Step generators tested.
-6. **Educational content**: Present and complete for all 7 sections.
-7. **Source files**: Exist for all 3 languages (TypeScript, Python, Java).
+6. **Educational content**: Present and complete for all 7 sections per `rules/algorithms.md`.
+7. **Source files**: Exist for all supported languages per `rules/algorithms.md`.
 8. **Line mappings**: Accurate per source file.
 9. **Non-persistence**: Input edits and grid edits are temporary.
 

.claude/agents/technical-writer.md

@@ -14,7 +14,7 @@ Review and improve all written content — educational algorithm explanations, p
 
 ## Review Areas
 
-1. **Educational content**: All 7 sections present (Overview, How It Works, Complexity, Best/Worst Case, Real-World Uses, Strengths/Limitations, When to Use)
+1. **Educational content**: All 7 sections present per `rules/algorithms.md`
 2. **ELI5 clarity**: Explanations use plain language, real-world analogies, and build from simple to complex
 3. **Documentation structure**: README.md and docs/ follow the structure defined in `.claude/rules/docs.md`
 4. **Contributor onboarding**: `docs/contributing.md` has clear step-by-step walkthrough for adding algorithms
@@ -32,7 +32,7 @@ Review and improve all written content — educational algorithm explanations, p
 - Never use jargon without first defining it in the same section
 - Educational content must be accurate — verify complexity claims against the actual implementation
 - Documentation updates must follow the trigger table in `.claude/rules/docs.md`
-- No references to AI, Claude, or automated generation in any documentation
+- No AI/Claude/assistant references per `rules/docs.md`
 
 ## Output Format
 

.claude/hooks/auto-plugin-mode.sh

@@ -1,21 +1,28 @@
 #!/usr/bin/env bash
 # SessionStart + branch creation hook: auto-switch plugins based on git branch prefix.
-# Reads plugin-profiles.json and updates settings.json enabledPlugins.
+# Reads plugin-profiles.json and updates settings.local.json enabledPlugins.
+# settings.json has all plugins enabled as baseline; this hook selectively
+# enables/disables per branch in the local override file.
 # Uses Node.js for JSON manipulation (no jq dependency).
 # Always exits 0 — never blocks session start.
 
 set -euo pipefail
 
 PROJECT_DIR="${CLAUDE_PROJECT_DIR:-$(git rev-parse --show-toplevel 2>/dev/null || pwd)}"
 PROFILES="$PROJECT_DIR/.claude/hooks/plugin-profiles.json"
-SETTINGS="$PROJECT_DIR/.claude/settings.json"
+SETTINGS="$PROJECT_DIR/.claude/settings.local.json"
 
-# Guard: profiles and settings must exist
-if [ ! -f "$PROFILES" ] || [ ! -f "$SETTINGS" ]; then
-  echo "WARN: plugin-profiles.json or settings.json missing — plugin auto-switching disabled" >&2
+# Guard: profiles must exist; settings.local.json is created if missing
+if [ ! -f "$PROFILES" ]; then
+  echo "WARN: plugin-profiles.json missing — plugin auto-switching disabled" >&2
   exit 0
 fi
 
+# Create settings.local.json if it doesn't exist
+if [ ! -f "$SETTINGS" ]; then
+  echo '{}' > "$SETTINGS"
+fi
+
 # Guard: Node.js must be available
 if ! command -v node &>/dev/null; then
   echo "WARN: node not found — plugin auto-switching disabled" >&2
@@ -33,7 +40,7 @@ if [ -n "${1:-}" ]; then
   BRANCH="$1"
 fi
 
-# Use Node.js to update settings.json atomically
+# Use Node.js to update settings.local.json atomically
 node -e "
 const fs = require('fs');
 

.claude/rules/algorithms.md

@@ -20,7 +20,7 @@ paths:
 
 - Real algorithm source file (pure implementation, no visualization logic)
 - Step generator using category-specific tracker
-- Multi-language source files: TypeScript, Python, Java
+- Multi-language source files: TypeScript, Python, Java, Rust, C++, Go
 - Unit tests for algorithm correctness
 - Unit tests for step generation
 - Pipeline story (`<AlgorithmName>Pipeline.stories.tsx`) co-located in the algorithm directory

.claude/rules/architecture.md

@@ -22,13 +22,21 @@
 
 ### Source Files
 
-- Real `.ts`, `.py`, `.java` files loaded via Vite `?raw` glob imports
+- Real `.ts`, `.py`, `.java`, `.rs`, `.cpp`, `.go` files loaded via Vite `?raw` glob imports
 - Source files are lintable, formattable artifacts - not embedded strings
 - Line mappings defined as static lookup tables in step generators
 
+### Directory Organization
+
+- Trackers: `src/trackers/<category>/` — grouped by algorithm category
+- Visualizers: `src/components/visualization/<category>/` — grouped by algorithm category
+- Algorithm tests + stories: `src/algorithms/<category>/<technique>/<algorithm>/__tests__/`
+- Source implementations: `src/algorithms/<category>/<technique>/<algorithm>/sources/`
+
 ### Adding New Algorithms
 
-1. Create `src/algorithms/<category>/<algorithm>/` directory
-2. Implement: index.ts, <algorithm>.ts, step-generator.ts, educational.ts, sources/
-3. Import in `src/algorithms/index.ts` barrel
-4. All UI works automatically via registry
+1. Create `src/algorithms/<category>/<technique>/<algorithm>/` directory
+2. Implement: index.ts, step-generator.ts, educational.ts, sources/ (6 languages)
+3. Add tests + pipeline story in `__tests__/`
+4. Import in `src/algorithms/index.ts` barrel
+5. All UI works automatically via registry

.claude/rules/storybook.md

@@ -6,8 +6,8 @@ paths:
 
 ## Storybook Rules
 
-- Pipeline stories (`*.Pipeline.stories.tsx`) live in algorithm directories, not `src/components/`
-- Component stories remain co-located with their components in `src/components/`
+- Pipeline stories (`*.Pipeline.stories.tsx`) live in the algorithm's `__tests__/` directory alongside test files
+- Component stories remain co-located with their components in `src/components/visualization/<category>/`
 - Every component gets at least one story per significant state variant
 - Visual regression via `@storybook/test-runner`
 - Test language switching in CodePanel stories

.claude/rules/testing.md

@@ -12,13 +12,14 @@ paths:
 ### Unit Tests (Vitest)
 
 - Unit tests must target actual algorithm implementations, not only step generators
-- Every algorithm needs: correctness tests (pure execute) + step generation tests
+- Every algorithm needs: correctness tests (pure execute) + step generation tests in `__tests__/`
 - Test tracker methods produce correct ExecutionStep with correct type/metrics
 - Test store slice state transitions for all actions
 - Test custom hooks with `renderHook`
 - Meaningful test variable names (no single chars)
 - Vitest uses the `projects` feature: `algorithms` project runs in `node` environment, `components` project runs in `jsdom`. This keeps total test time ~20 seconds and removes the need for manual timeout hooks in `test-setup.ts`
-- CI shards unit tests 8 ways; aggregation job is named **Unit Tests Status**
+- CI shards unit tests 12 ways; aggregation job is named **Unit Tests Status**
+- Language tests (Python, Java, Rust, C++, Go) sharded separately per language in CI
 
 ### Coverage Thresholds
 
@@ -47,4 +48,4 @@ paths:
 - Step count and step types for known inputs
 - Final visual state matches expected
 - Educational content is non-empty for all sections
-- Source files exist for all supported languages
+- Source files exist for all supported languages (TypeScript, Python, Java, Rust, C++, Go)

.claude/rules/ui-ux.md

@@ -29,7 +29,7 @@ paths:
 ### Code Panel
 
 - Monaco editor in read-only mode (default)
-- Language tabs: TypeScript, Python, Java
+- Language tabs: TypeScript, Python, Java, Rust, C++, Go
 - Synchronized line highlighting per current step
 - Temporary editable mode (non-persistent)
 

.claude/settings.json

@@ -1,32 +1,6 @@
 {
   "permissions": {
-    "allow": [
-      "Bash(node /tmp/check_quick.mjs)",
-      "Bash(node /tmp/check_bitonic.mjs)",
-      "Bash(node /tmp/check_sleep.mjs)",
-      "Bash(node /tmp/check_sleep2.mjs)",
-      "Bash(node /tmp/check_cartesian.mjs)",
-      "Bash(node /tmp/check_library.mjs)",
-      "Bash(node /tmp/check_radix_msd.mjs)",
-      "Bash(node /tmp/check_counting.mjs)",
-      "Bash(node /tmp/check_bead.mjs)",
-      "Bash(node /tmp/check_bead2.mjs)",
-      "Bash(node /tmp/check_smooth_fixed.mjs)",
-      "Bash(node /tmp/check_radix_fixed.mjs)",
-      "Bash(node /tmp/check_radix_fixed2.mjs)",
-      "Bash(node /tmp/check_radix_fixed3.mjs)",
-      "Bash(node /tmp/verify_sleep.mjs)",
-      "Bash(node /tmp/check_library2.mjs)",
-      "Bash(killall -9 node)",
-      "Bash(tee /tmp/e2e_run2.txt)",
-      "Bash(echo \"DONE_EXIT:$?\")",
-      "Bash(git -C /Users/springfield/dev/algo_flow add -A)",
-      "Bash(git -C /Users/springfield/dev/algo_flow push)",
-      "Bash(xargs -I {} sh -c 'echo \"=== {} ===\" && find {} -name \"index.ts\" -type f | wc -l')",
-      "Bash(npx playwright:*)",
-      "Bash(curl -s http://localhost:5174)",
-      "Bash(curl -s -o /dev/null -w \"%{http_code}\" http://localhost:5174/)"
-    ]
+    "allow": []
   },
   "hooks": {
     "PreToolUse": [
@@ -147,15 +121,15 @@
     "github@claude-plugins-official": true,
     "code-review@claude-plugins-official": true,
     "pr-review-toolkit@claude-plugins-official": true,
-    "frontend-design@claude-plugins-official": false,
-    "figma@claude-plugins-official": false,
-    "playground@claude-plugins-official": false,
-    "playwright@claude-plugins-official": false,
-    "security-guidance@claude-plugins-official": false,
-    "code-simplifier@claude-plugins-official": false,
-    "claude-md-management@claude-plugins-official": false,
-    "skill-creator@claude-plugins-official": false,
-    "claude-code-setup@claude-plugins-official": false,
-    "ralph-loop@claude-plugins-official": false
+    "frontend-design@claude-plugins-official": true,
+    "figma@claude-plugins-official": true,
+    "playground@claude-plugins-official": true,
+    "playwright@claude-plugins-official": true,
+    "security-guidance@claude-plugins-official": true,
+    "code-simplifier@claude-plugins-official": true,
+    "claude-md-management@claude-plugins-official": true,
+    "skill-creator@claude-plugins-official": true,
+    "claude-code-setup@claude-plugins-official": true,
+    "ralph-loop@claude-plugins-official": true
   }
 }

.claude/skills/debugging/SKILL.md

@@ -58,7 +58,7 @@ Symptoms: empty code panel, wrong language, misaligned highlighting
 1. Verify Vite `?raw` glob import returns content (not `{}`)
 2. Check file paths in glob pattern — are they statically analyzable?
 3. Verify `LineHighlight` mappings per language match actual source file line numbers
-4. Check all 3 languages have source files AND line mappings
+4. Check all 6 languages have source files AND line mappings
 5. Test language tab switching — does `useAlgorithmSource` update correctly?
 
 ### Grid Editing Bugs (Pathfinding)