@redaksjon/brennpunkt 0.0.2 → 0.1.0

package/guide/ai-integration.md

@@ -121,6 +121,48 @@ All tools read EXISTING coverage data (lcov.info) — they do NOT run tests.
 | `brennpunkt_get_file_coverage` | Detailed coverage for a specific file |
 | `brennpunkt_estimate_impact` | "If I test these files, will I hit 90%?" |
 
+### Available MCP Resources (NEW in v0.1.0)
+
+Resources provide direct access to coverage data without tool calls, enabling AI assistants to read and analyze data more efficiently.
+
+| Resource URI | Purpose |
+|--------------|---------|
+| `brennpunkt://coverage/{projectPath}` | Full coverage data as JSON for complex analysis |
+| `brennpunkt://file/{projectPath}/{filePath}` | Detailed single-file coverage |
+| `brennpunkt://priorities?project={path}&top={n}` | Pre-ranked priority list with filtering |
+| `brennpunkt://config/{projectPath}` | Project configuration (yaml or defaults) |
+| `brennpunkt://quick-wins?project={path}` | Small files with high impact potential |
+
+**Why Resources?** Resources allow AI to:
+- Read coverage data once and reason about it multiple times
+- Perform complex analysis like "Compare branch coverage across all auth files"
+- Access cacheable datasets that change infrequently
+- Fetch multiple resources in parallel
+
+### Available MCP Prompts (NEW in v0.1.0)
+
+Prompts are workflow templates that guide coverage improvement, transforming brennpunkt from a data provider to a coverage improvement partner.
+
+| Prompt | Purpose | Key Arguments |
+|--------|---------|---------------|
+| `improve_coverage` | Complete workflow to reach target percentage | projectPath, targetPercentage, focusMetric |
+| `analyze_gaps` | Understand patterns in coverage gaps | projectPath, targetPercentage |
+| `quick_wins_workflow` | Find fast paths to improvement | projectPath, timeConstraint |
+| `coverage_review` | Detailed review with test suggestions | projectPath, filePattern |
+
+**Example Prompt Usage:**
+
+```
+Human: I need to get to 90% coverage
+
+AI: [invokes improve_coverage prompt with projectPath and targetPercentage=90]
+AI: I'll help you reach 90% coverage. Let me check your current state...
+AI: [calls brennpunkt_coverage_summary]
+AI: You're at 82%, need 8% improvement...
+AI: [calls brennpunkt_get_priorities]
+AI: Here are the top 3 files to test: ...
+```
+
 ### Setting Up MCP
 
 #### 1. Install Brennpunkt
@@ -131,26 +173,26 @@ npm install -g @redaksjon/brennpunkt
 
 #### 2. Configure MCP Client (One-Time Setup)
 
-For **Cursor**, add to your MCP configuration:
+Add to your MCP configuration (**Cursor** or **Claude Desktop**):
 
 ```json
 {
   "mcpServers": {
     "brennpunkt": {
-      "command": "brennpunkt-mcp"
+      "command": "npx",
+      "args": ["-y", "-p", "@redaksjon/brennpunkt", "brennpunkt-mcp"]
     }
   }
 }
 ```
 
-For **Claude Desktop**, add to `claude_desktop_config.json`:
+Or if installed globally (`npm install -g @redaksjon/brennpunkt`):
 
 ```json
 {
   "mcpServers": {
     "brennpunkt": {
-      "command": "npx",
-      "args": ["@redaksjon/brennpunkt-mcp"]
+      "command": "brennpunkt-mcp"
     }
   }
 }

package/guide/index.md

@@ -100,7 +100,7 @@ If you're an AI helping someone improve test coverage:
 4. **Use JSON output**: Parse the structured data for precise guidance
 5. **Respect project config**: Projects may have custom `brennpunkt.yaml` with different weights
 
-> **MCP Server**: Run `brennpunkt-mcp` to start an MCP server that allows AI tools to query coverage data directly without running tests. See [AI Integration](./ai-integration.md) for details.
+> **MCP Server**: Use `npx -p @redaksjon/brennpunkt brennpunkt-mcp` to start an MCP server that allows AI tools to query coverage data directly without running tests. See [AI Integration](./ai-integration.md) for details.
 
 ### Cursor Rule (Recommended)
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@redaksjon/brennpunkt",
-  "version": "0.0.2",
+  "version": "0.1.0",
   "description": "Coverage priority analyzer - identify where to focus testing efforts",
   "main": "dist/index.js",
   "type": "module",
@@ -31,8 +31,13 @@
     "lint": "eslint . --ext .ts",
     "lint:fix": "eslint . --ext .ts --fix",
     "clean": "rm -rf dist",
-    "precommit": "npm run lint && npm run test",
-    "prepublishOnly": "npm run clean && npm run build"
+    "precommit": "npm run lint && npm run test && npm run mcp:test",
+    "prepublishOnly": "npm run clean && npm run build",
+    "mcp:build": "npm run build && echo 'MCP server built'",
+    "mcp:inspect": "npm run build && npx @modelcontextprotocol/inspector dist/mcp/server.js",
+    "mcp:inspect:dev": "nodemon --watch src --ext ts --exec 'npm run mcp:inspect'",
+    "mcp:test": "node scripts/test-mcp-compliance.js",
+    "mcp:start": "node dist/mcp/server.js"
   },
   "keywords": [
     "coverage",
@@ -59,9 +64,10 @@
   "devDependencies": {
     "@eslint/eslintrc": "^3.3.1",
     "@eslint/js": "^9.39.1",
+    "@modelcontextprotocol/inspector": "^0.18.0",
     "@rollup/plugin-replace": "^6.0.3",
     "@swc/core": "^1.15.1",
-    "@types/node": "^24.10.1",
+    "@types/node": "^25.0.9",
     "@typescript-eslint/eslint-plugin": "^8.46.4",
     "@typescript-eslint/parser": "^8.46.4",
     "@vitest/coverage-v8": "^4.0.8",