First Scan Report

The First Scan Report is the initial analysis of your entire codebase --- not just a PR diff, but a full scan of all TypeScript and JavaScript files in the project. This is the "shock report" that reveals the current state of technical debt your team has accumulated.

Every new project connected to Radar receives a First Scan Report automatically. It is designed to give you an honest, unflinching view of where you stand, and a clear path to improvement.

What to Expect

First scan results are often surprising. Codebases that "feel clean" to the team frequently show dozens or hundreds of violations that have accumulated over months or years of incremental changes. This is normal.

The report is not a judgment --- it is a baseline. The value of Radar is not the first scan score; it is the trend that follows. Every PR that passes the gate brings the score down.

Report Sections

Summary Gauges

Four circular gauges at the top of the report provide an at-a-glance summary:

Gauge	What It Shows
Overall Debt Score	A 0--100 score aggregating all violations. 0 is pristine, 100 is critical debt across every category. The gauge needle and color indicate severity: green (0--30), yellow (31--60), red (61--100).
Architecture Health	Percentage of modules with clean architecture (no boundary violations, no circular dependencies).
Runtime Safety	Percentage of request handlers free from event-loop-blocking patterns.
Performance Risk	Count of unbounded queries and N+1 patterns, weighted by table volume.

Top Violations

A prioritized list of the 10 most impactful violations in the codebase, ranked by a combination of severity and blast radius (how many files are affected).

Each violation entry shows:

• Rank (1--10)
• Rule ID and category
• Severity badge (Critical, Warning)
• File path and line number of the primary occurrence
• Occurrence count --- how many times this same violation appears across the codebase
• Description --- plain-language explanation of the risk
• Fix guidance --- concrete steps to resolve the violation

Tip: Start with violation #1 and work down. The ranking prioritizes violations that, when fixed, produce the largest score improvement.

Category Breakdown

A detailed breakdown of violations by category:

Architecture Debt:

• Total layer boundary violations
• Total circular dependencies (with cycle paths listed)
• Module boundary violations
• Forbidden framework imports in domain layers

Runtime Risk Debt:

• Sync I/O in request handlers (count and file list)
• Sync crypto operations
• ReDoS-vulnerable patterns
• Unbounded JSON/Buffer operations

Performance Debt:

• Unbounded queries by table volume (S/M/L/XL/XXL)
• N+1 query patterns
• Missing pagination on endpoints
• Fetch-all-filter-in-memory patterns

Reliability Debt:

• Fire-and-forget promises
• Missing timeouts on HTTP calls
• Empty catch blocks
• Missing error handling in use cases

Maintainability Debt:

• Functions exceeding complexity threshold (listed with their cyclomatic complexity)
• Files exceeding line count threshold

Each category section is collapsible. Expand a category to see the full list of violations with file paths and line numbers.

Copy for AI

The Copy for AI button is one of the most powerful features of the First Scan Report. It generates a formatted text version of the report optimized for pasting into AI coding tools.

What It Copies

The clipboard content includes:

## Technical Debt Radar - First Scan Report
Project: my-api
Scanned: 2026-03-18
Files analyzed: 342
Total violations: 47 (12 critical, 28 warnings, 7 info)

### Top 10 Violations (fix these first)

1. [CRITICAL] layer-boundary-violation
   File: src/orders/domain/order.service.ts:14
   Import: @prisma/client
   Fix: Move database access to src/orders/infrastructure/order.repository.ts

2. [CRITICAL] sync-fs-in-handler
   File: src/api/controllers/export.controller.ts:28
   Pattern: fs.readFileSync('./template.html')
   Fix: Replace with await fs.promises.readFile('./template.html', 'utf-8')

...

### All Violations by Category
[Full categorized list with file paths and line numbers]

### Fix Priority
1. Fix all Critical violations first (blocks merges)
2. Address XL/XXL performance violations (blocks merges)
3. Reduce Warning count by 50% in next sprint

How to Use It

The AI feedback loop:

1. Click Copy for AI on the First Scan Report
2. Open your AI coding tool (Claude Code, Cursor, Windsurf, or any AI assistant)
3. Paste the report and ask the AI to fix the violations
4. The AI reads the file paths, line numbers, and fix instructions, then generates fixes
5. Run radar scan locally to verify the fixes
6. Push the changes --- Radar verifies on the PR

This workflow can resolve 50--80% of first scan violations in a single session, depending on codebase size and violation types.

Tip: For large codebases with many violations, paste only the top 10 violations first. Fix those, re-scan, then paste the next batch. This prevents the AI from being overwhelmed with too many changes at once.

Using the Report to Prioritize Fixes

Week 1: Critical Blockers

Focus exclusively on Critical violations because these block every future PR that touches the affected files:

1. Architecture violations (layer boundaries, circular deps)
2. Runtime risk violations (sync I/O in handlers)
3. N+1 queries (always critical)

Week 2: High-Volume Performance

Address performance violations on your largest tables:

1. Unbounded queries on XL and XXL tables
2. Missing pagination on public-facing endpoints
3. Fetch-all-filter-in-memory patterns

Week 3: Reliability and Maintainability

Clean up reliability patterns and start reducing complexity:

1. Fire-and-forget promises (add await or .catch())
2. Missing timeouts on HTTP calls
3. Split the highest-complexity functions

Ongoing: Gate Enforcement

After the initial cleanup, Radar's PR gate prevents new violations from being introduced. The debt score should trend downward over time as existing violations are fixed and no new ones are added.

Re-Running the First Scan

The First Scan Report is generated once when a project is first connected. To generate an updated full-project scan:

1. Go to Settings > Project
2. Click Re-run Full Scan

This creates a new First Scan Report with current results, useful after a major refactoring effort to measure progress against the original baseline.

The original First Scan Report is preserved for comparison. You can view both the original and the most recent full scan side by side.