@vyuhlabs/dxkit 0.1.0 → 0.1.1

package/README.md

@@ -0,0 +1,241 @@
+# @vyuhlabs/dxkit
+
+AI-native developer experience toolkit for any repository. Adds Claude Code agents, skills, commands, and quality hooks to existing projects in seconds.
+
+## Quick Start
+
+```bash
+# Auto-detect your stack and set up everything
+npx @vyuhlabs/dxkit init --detect
+
+# Interactive mode (prompts for config)
+npx @vyuhlabs/dxkit init
+
+# Full mode (DX + quality + hooks + CI)
+npx @vyuhlabs/dxkit init --full --yes
+```
+
+## What It Does
+
+Running `init` auto-detects your tech stack and generates a complete `.claude/` directory with:
+
+```
+.claude/
+  settings.json          # Permissions, deny list, hooks
+  agents/                # Active agents (auto-trigger)
+    knowledge-bot.md     # Answers codebase questions
+    onboarding.md        # Interactive onboarding buddy
+    quality-reviewer.md  # Reviews code before committing
+  agents-available/      # Dormant agents (activate on demand)
+    codebase-explorer.md # Deep architecture analysis
+    code-reviewer.md     # PR review and security audit
+    test-writer.md       # Writes tests for existing code
+    debugger.md          # Root cause analysis
+  commands/              # Slash commands
+    ask.md               # /ask — Query the codebase
+    quality.md           # /quality — Run linters
+    test.md              # /test — Run tests with coverage
+    check.md             # /check — Full pre-commit validation
+    fix.md               # /fix — Auto-fix issues
+    learn.md             # /learn — Capture gotchas/conventions
+    onboarding.md        # /onboarding — Start onboarding
+    help.md              # /help — List everything
+    session-start.md     # /session-start — Begin dev session
+    session-end.md       # /session-end — Checkpoint + evolve skills
+    ...
+  skills/                # Domain knowledge
+    codebase/            # Auto-generated architecture overview
+    learned/             # Evolving gotchas, conventions, deny list
+    quality/             # Quality check patterns
+    ...
+  rules/                 # Path-scoped rules (per language)
+CLAUDE.md                # Main context file for Claude Code
+.ai/
+  sessions/              # Session checkpoints
+```
+
+## Supported Languages
+
+| Language | Detection | Linters | Tests | Coverage |
+|---|---|---|---|---|
+| Node.js / TypeScript | `package.json` | ESLint, Prettier, tsc | Jest | jest --coverage |
+| Python | `pyproject.toml`, `setup.py` | ruff, mypy | pytest | pytest-cov |
+| Go | `go.mod` | golangci-lint, go vet | go test | go cover |
+| C# | `*.csproj`, `*.sln` | dotnet format, Roslyn | dotnet test | coverlet |
+| Rust | `Cargo.toml` | clippy, rustfmt | cargo test | tarpaulin |
+
+## Supported Tools
+
+Auto-detected and integrated when present:
+
+- **Google Cloud** (gcloud) — SDK commands, security rules
+- **Infisical** — Secrets management, never-leak rules
+- **Pulumi** — IaC with preview-before-apply safety
+- **Docker** — Container commands
+
+## Commands
+
+After init, these slash commands are available in Claude Code:
+
+### Development Workflow
+| Command | Description |
+|---|---|
+| `/session-start` | Start an AI-assisted dev session |
+| `/session-end` | End session, create checkpoint, evolve skills |
+| `/ask <question>` | Ask about the codebase (delegates to knowledge-bot) |
+| `/learn` | Capture a gotcha, convention, or thing to avoid |
+
+### Quality & Testing
+| Command | Description |
+|---|---|
+| `/quality` | Run language-specific linters and type checks |
+| `/test` | Run tests with coverage enforcement |
+| `/check` | Full pre-commit validation (quality + tests + AI review) |
+| `/fix` | Auto-fix formatting and lint issues |
+| `/setup-hooks` | Install git pre-commit/pre-push hooks |
+
+### Exploration & Onboarding
+| Command | Description |
+|---|---|
+| `/onboarding` | Interactive onboarding for new developers |
+| `/explore-codebase` | Deep architecture exploration |
+| `/help` | List all commands and agents |
+
+### Operations
+| Command | Description |
+|---|---|
+| `/fix-issue <number>` | Investigate and fix a GitHub issue |
+| `/doctor` | Diagnose environment issues |
+| `/enable-agent <name>` | Activate a dormant agent |
+
+## Agents
+
+### Active by Default
+These agents auto-trigger when Claude detects a matching question:
+
+- **knowledge-bot** — "How does auth work?" "Where are payments handled?"
+- **onboarding** — "I'm new, help me get started" "What does this project do?"
+- **quality-reviewer** — "Review my changes" "Check quality before I commit"
+
+### Dormant (activate with `/enable-agent`)
+- **codebase-explorer** — Deep architecture analysis, generates documentation
+- **code-reviewer** — PR review and security audit (read-only)
+- **test-writer** — Writes tests for existing code
+- **debugger** — Systematic root cause analysis
+
+## Learning System
+
+DXKit includes a continuous learning system that improves over time:
+
+1. **Stop Hook** — After each conversation, Claude is reminded to capture learnings
+2. **`/learn` command** — Explicitly save gotchas, conventions, or things to avoid
+3. **`/session-end`** — Creates checkpoint and evolves skill files
+4. **Evolving files** — These files are append-only and never overwritten:
+   - `.claude/skills/learned/references/gotchas.md`
+   - `.claude/skills/learned/references/conventions.md`
+   - `.claude/skills/learned/references/deny-recommendations.md`
+
+## Coverage Enforcement
+
+Tests run with a minimum coverage threshold (default: 80%). Configure at init time or in `.vyuh-dxkit.json`.
+
+## CLI Reference
+
+```bash
+# Initialize (auto-detect stack)
+npx @vyuhlabs/dxkit init --detect
+
+# Initialize (interactive)
+npx @vyuhlabs/dxkit init
+
+# Initialize (everything, no prompts)
+npx @vyuhlabs/dxkit init --full --yes
+
+# Re-generate after config changes (preserves evolved files)
+npx @vyuhlabs/dxkit update
+
+# Re-generate with codebase rescan
+npx @vyuhlabs/dxkit update --rescan
+
+# Verify setup
+npx @vyuhlabs/dxkit doctor
+```
+
+### Init Options
+
+| Flag | Description |
+|---|---|
+| `--detect` | Auto-detect stack, minimal prompts |
+| `--yes` | Accept all defaults |
+| `--dx-only` | Just `.claude/` + `CLAUDE.md` (default) |
+| `--full` | Everything: DX + quality + hooks + CI |
+| `--force` | Overwrite existing files (except evolved) |
+| `--name <n>` | Override project name |
+| `--no-scan` | Skip codebase analysis |
+
+## Example: Adding to an Existing Node.js Project
+
+```bash
+cd my-express-app
+npx @vyuhlabs/dxkit init --detect --yes
+```
+
+This generates:
+- `CLAUDE.md` with Node.js-specific context
+- `.claude/commands/quality.md` with `npx eslint .` and `npx tsc --noEmit`
+- `.claude/commands/test.md` with `npx jest --coverage` (80% threshold)
+- `.claude/rules/` with Node.js conventions
+- `.claude/skills/codebase/SKILL.md` with auto-scanned architecture overview
+- 3 active agents + 4 dormant agents
+
+Then in Claude Code:
+```
+/help                              # See everything available
+/ask How does the auth middleware work?  # Ask about the codebase
+/quality                           # Run ESLint + TypeScript checks
+/onboarding                        # Get oriented as a new developer
+```
+
+## Example: Adding to a Python + Go Monorepo
+
+```bash
+cd my-monorepo
+npx @vyuhlabs/dxkit init --detect --yes
+```
+
+Detects both languages and generates:
+- Quality commands with ruff + mypy (Python) and golangci-lint (Go)
+- Test commands with pytest + go test
+- Path-scoped rules for both `.py` and `.go` files
+- Language-specific coverage enforcement
+
+## How It Works
+
+1. **Detection** — Scans for config files (`package.json`, `go.mod`, `pyproject.toml`, etc.) to determine languages and tools
+2. **Template Processing** — Processes `.md.template` files through a conditional template engine, stripping irrelevant language sections
+3. **Codebase Scanning** — Analyzes source files to find entry points, API routes, test patterns, and conventions
+4. **Generation** — Writes files non-destructively (never overwrites without `--force`, evolving files always preserved)
+5. **Manifest** — Saves state to `.vyuh-dxkit.json` for `update` and `doctor` commands
+
+## Non-Destructive by Design
+
+- Files that already exist are **skipped** (use `--force` to overwrite)
+- Evolving files (`gotchas.md`, `conventions.md`, `deny-recommendations.md`) are **never overwritten**, even with `--force`
+- The manifest (`.vyuh-dxkit.json`) tracks what was generated for safe updates
+
+## AI-Assisted Development Workflow
+
+```
+/session-start                          # Load context, plan work
+git checkout -b feature/my-feature      # Create branch
+/ask How does the payment flow work?    # Understand the code
+# ... develop with full context ...
+/quality                                # Lint + AI review
+/test                                   # Tests with coverage
+git add -A && git commit                # Commit
+/session-end                            # Checkpoint + evolve skills
+```
+
+## License
+
+MIT

package/dist/constants.d.ts

@@ -1,5 +1,5 @@
 import { ResolvedConfig } from './types';
-export declare const VERSION = "0.1.0";
+export declare const VERSION = "0.1.1";
 export declare const DEFAULT_VERSIONS: {
     python: string;
     go: string;

package/dist/constants.js

@@ -3,7 +3,7 @@ Object.defineProperty(exports, "__esModule", { value: true });
 exports.EVOLVING_FILES = exports.DEFAULT_COVERAGE = exports.DEFAULT_VERSIONS = exports.VERSION = void 0;
 exports.buildVariables = buildVariables;
 exports.buildConditions = buildConditions;
-exports.VERSION = '0.1.0';
+exports.VERSION = '0.1.1';
 exports.DEFAULT_VERSIONS = {
     python: '3.12',
     go: '1.24.0',

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@vyuhlabs/dxkit",
-  "version": "0.1.0",
+  "version": "0.1.1",
   "description": "AI-native developer experience toolkit for any repository",
   "license": "MIT",
   "bin": {