Scout Analysis

After a Scout mission completes, the analysis view presents all findings, the generated app map, and detailed exploration telemetry. This page covers how to interpret and act on Scout results.

Deep Exploration Results

When Scout finishes a Deep Exploration mission, the results panel shows:

Section	Content
Pages Discovered	Every unique URL visited, with status codes and load times
Interactions Performed	Forms filled, buttons clicked, modals opened, dropdowns selected
State Transitions	Navigation paths between pages, including back-button and redirect behavior
Console Output	JavaScript errors, warnings, and uncaught exceptions captured per page
Screenshots	Timestamped captures at each major navigation point

Bug Hunter Findings

Bug Hunter missions produce a structured findings list. Each finding includes:

• Severity — Critical, High, Medium, Low, or Info.
• Category — Functional bug, UI glitch, console error, crash, or performance regression.
• Reproduction steps — exact sequence of actions Scout performed to trigger the issue.
• Evidence — screenshot at the moment of failure plus DOM snapshot.

TIP

Click any finding to replay the exact navigation sequence in Scout's timeline view. This lets you see what happened step-by-step leading up to the bug.

App Map Generation

During exploration, Scout builds a directed graph of your application's navigation structure:

Home --> Dashboard --> Settings
  |        |            └--> Profile
  |        └--> Reports --> Export
  └--> Login --> Forgot Password

The app map reveals:

• Orphaned pages — routes not reachable from the main navigation.
• Dead ends — pages with no outbound links or actions.
• Deep nesting — workflows that require many clicks to reach.
• Circular paths — navigation loops.

INFO

The app map is stored in .xyva/agent-memory/ within your project and updates with each new Scout run. Previous maps are retained for comparison.

Autonomous Navigation

Scout decides where to go next using a priority queue:

1. Unvisited links are ranked highest.
2. Forms and inputs are prioritized over static content.
3. Error-prone areas (pages with previous console errors) are revisited with different inputs.
4. Depth limits prevent infinite crawling — configurable in Expert Mode settings.

Scout never follows external links or navigates away from the target domain unless explicitly configured to do so.

Exporting Results

• JSON — machine-readable export of all findings and the app map for CI integration.
• PDF — formatted report with screenshots, suitable for stakeholder review.
• YouTrack — direct issue creation for confirmed bugs (requires YouTrack integration in Settings).

WARNING

Scout analysis data can be large for complex applications. Runs against 100+ page apps may produce 50MB+ of screenshots and DOM snapshots. Configure screenshot quality in Settings to manage storage.

Related

• Scout Overview
• Swarm QA Findings
• Swarm QA Overview