Browser Forensics Hands-on Labs¶
Choose the Interactive Lab to practise BrowsingHistoryView in your browser, or the Browser Forensics Full Lab to trace real training activity across history, download records and a saved file in a Windows VM.
- No installation
- Guided case
- Real tools
- Three activities
Interactive Lab¶
BrowsingHistoryView Interactive Lab
Enable JavaScript for this simulation. The Full Lab below provides the real-tool workflow.
Full Lab¶
Browser Forensics Full Lab
Case BRH-01 / follow the download lead
Two profiles. One download. A finding you can defend.
An employee reports a support-page visit followed by an unexpected download. Recreate the activity with harmless local pages. Use BrowsingHistoryView to trace the visits, BrowserDownloadsView to check the download record, and the saved file to check the bytes. Hand over a short finding another analyst can follow.
Tick a checkpoint after saving its evidence. Progress is self reported and saved in this browser; the page cannot check your VM. Keep your case files separately.
Investigation brief¶
Two labelled history exports, a download record, a file comparison and a four-sentence handover.
Real case note: History records can support a visit. They do not establish download completion or execution. Two viewers of the same database are not independent evidence.
Before you start¶
Need: a snapshotted Windows VM, Python 3, Chrome, BrowsingHistoryView, BrowserDownloadsView and separate case storage. Use two fresh, unsynchronised Chrome profiles with no personal data. No administrator session is needed for your own training profiles.
Plan: allow 45–60 minutes after setup. Complete the three activities in order; each ends with a save point. Searches, cookies and cache are optional extensions.
Download a blank case worksheet to your notes folder. Record your own results; no browser evidence or expected timestamps are supplied.
Need the tools or a simpler starting point?
Use the download and interface guidance for BrowsingHistoryView and BrowserDownloadsView. Record the exact builds you use. If VM setup is new to you, try the Interactive Lab above first. It needs no installation.
Activity 1: Prepare and capture the activity¶
1. Set up a fresh case and the local site
Create a fresh case folder on separate storage, for example E:\Cases\BrowserTraining-01, with sources, working, exports and notes subfolders. Retain existing attempts; use a new case path rather than overwriting one. Record host, OS, browser/tool versions, time zone/UTC offset, disk identities and snapshot.
Save and review serve_lab.py. Run it in Command Prompt on the disposable VM, using your actual path:
The helper binds only to loopback. It creates a harmless text attachment, an actual source manifest and UTC request log. It refuses an existing site folder; resolve a startup failure before browsing. It does not create browser history, an executable or preset forensic results.
In Chrome's profile menu, add two profiles without signing in. Label them Training A and Training B. In each, open chrome://version and record Profile Path. The folder names may be Default or Profile 1; use the displayed paths, not the labels, when collecting. A profile name does not identify who acted.
Look for: a running local site, fresh case folders and saved context. No browser evidence image is supplied: you generate the activity yourself.
2. Generate two different activity paths and preserve them
In Training A, open the printed local URL, visit Help and click Download training text once. Save it in a dedicated training folder and note its actual path and action time. In Training B, open the same site and visit Case notes instead; do not download the file there. Record both profiles' actions in the worksheet.
Close both training browsers cleanly. In Task Manager, confirm Chrome has exited before copying; background Chrome processes can keep databases open. Copy each entire Profile Path folder to sources\A and sources\B, retaining its contents and subfolders. These destinations should contain History directly. Preserve any History-wal, History-shm or History-journal files present alongside it. Never copy a changing database and assume it is consistent.
This is a disposable-VM collection exercise. In a real incident, choose a validated acquisition method before changing a live system. Closing a browser can change session evidence.
Make separate working copies and compare corresponding files by relative path, count, byte size and full SHA-256. Keep manifests outside the directories being hashed. Resolve mismatches before parsing. Retain an unchanged copy of the harmless downloaded file with its size and hash. A hash identifies retained bytes; it does not establish acquisition completeness.
How do I check the copied files?
Copy sources\A to working\A, and repeat for B. In Windows PowerShell, set your real case path and save a list of every file's relative path, size and SHA-256:
$browserCase = 'E:\Cases\BrowserTraining-01'
foreach ($set in 'sources', 'working') {
$browserRoot = (Resolve-Path -LiteralPath "$browserCase\$set" -ErrorAction Stop).Path
Get-ChildItem -LiteralPath $browserRoot -Recurse -File -Force -ErrorAction Stop |
ForEach-Object {
[pscustomobject]@{
RelativePath = $_.FullName.Substring($browserRoot.Length + 1)
Bytes = $_.Length
SHA256 = (Get-FileHash -LiteralPath $_.FullName -Algorithm SHA256 -ErrorAction Stop).Hash
}
} | Export-Csv -LiteralPath "$browserCase\notes\$set-sha256.csv" -NoTypeInformation
}
Compare-Object (Import-Csv -LiteralPath "$browserCase\notes\sources-sha256.csv") `
(Import-Csv -LiteralPath "$browserCase\notes\working-sha256.csv") `
-Property RelativePath, Bytes, SHA256
No comparison rows means the saved lists match. Check both CSVs contain files for A and B; an empty list is not a successful collection. If there is an error or difference, keep the output, resolve the copy issue and create new comparison lists. Save this check before opening the working copies in a tool.
Look for: two consistently acquired, separately labelled profile sets, checked working copies and actual action notes. Browser output, hashes and times must come from your run.
Next: find the support-page trail in the copied history.
Continue to the history check
Activity 2: Examine the copied history¶
Before opening the tool, note which profile should contain Help and which should contain Case notes. Then test that prediction against the copied records. Keep an unexpected or missing result in your notes.
3. Load the working sources in BrowsingHistoryView
Open BrowsingHistoryView. Press F9 and choose Load history from the specified history files. Select Chrome and enter your actual working\A\History path in the Chrome history-file field. Set the time range to cover your actions, then load. This avoids depending on the analyst's Windows profile or registry. Save the source path, range and displayed time handling in the worksheet.
Examine each profile separately; repeat with working\B\History. Retain History File and Record ID where available. A direct file load may leave user/profile fields blank; your original-path-to-A/B mapping supplies that context. For interface screenshots, see BrowsingHistoryView.
Look for: records you can trace to a known training action. If absent, check input path, format support, filters, time conversion, companion files and load errors. An empty table does not prove no visit occurred.
4. Filter, inspect and export the relevant visits
Use View → Use Quick Filter (Ctrl+Q) to search for the actual local host or a distinctive page path. Inspect the source/profile and visit fields for each relevant row. Use the properties view; do not open a case URL in a networked analyst browser. Double-click behaviour is configurable, so select the properties action appropriate to your build.
Find Training A's Help-page trail and Training B's Case notes trail. Note any extra records. Do not merge same-URL records across profiles into one actor story. A second viewer of the same database can cross-check parsing; it is not independent activity evidence.
Select relevant rows and use File → Save Selected Items to export CSV with headers to exports. Record whether the export contains selected or all records, exact filters, parser version and raw/displayed time context. Keep original values when normalising to UTC.
Save A and B separately, for example A-visits.csv and B-visits.csv. Open each as a text table and check the headers and relevant rows were saved. Record the original profile path and copied History path beside each export; the filename alone is not enough.
Look for: a source-labelled export with actual observed records. No row count, record number or timestamp is prescribed by the exercise.
Next: test the download lead against the saved file.
Continue to the file comparisonActivity 3: Correlate and report¶
5. Check the download against independent evidence
In BrowserDownloadsView, press F9 and choose Load downloads from the specified Web browser profile folders. Enter your actual working\A folder in the Chrome profile field. Find the attachment record, inspect its available URL, target path, state, byte count and start/end times, then select it and use Ctrl+S to save CSV with headers. Record fields your build does not expose. Check B separately as a comparison; do not assume it must be empty.
Compare the retained harmless attachment with site\source-manifest.json by content, size and full SHA-256. In PowerShell, use Get-Item -LiteralPath 'your retained file path' for size and Get-FileHash -Algorithm SHA256 -LiteralPath 'your retained file path' for the hash. Save the actual values. Different paths are expected. Keep a mismatch and investigate wrong record/output selection or changes; do not edit the manifest to force a match. Use Download Analysis for the focused workflow.
Compare the request log with your action times as training context. Both tools can read the same Chrome History database: this is a comparison of related records, not two independent acquisitions. The retained attachment supplies the file bytes. A server request alone does not prove that the browser saved a complete file.
Look for: a documented browser-to-file comparison, or an explained gap. Do not run an executable recovered in a real case to test it; this lab supplies only harmless text.
6. Write and review the handover
Write four short sentences in the worksheet: what you observed, which saved records and file comparison support it, what remains unknown, and what you would check next. Use your own paths, times and results. Preserve a failed or inconclusive result when that is what your sources support.
Include profile/source, row identifier, URLs, time meanings, export paths, settings and full hashes where relevant. Record missing records, unsupported formats, sync/background possibilities and clock uncertainty. A visit or profile label does not prove who clicked, what they intended or that software ran.
Ask a peer to trace one observation from the note back to its source, or repeat that trace yourself. Stop the helper with Ctrl+C, then retain the final request log and source manifest. Check that the saved case package is outside VM rollback storage before restoring the snapshot. Retain failed attempts and record cleanup.
Look for: a reproducible finding with a useful next evidence check, such as endpoint process records for an execution question.
Finish: review the saved evidence package below.
Review your handover- Sources: labelled profile copies, original paths, working-copy checks and the retained attachment.
- Context: worksheet, actual action times, time zone, versions, final request log and source manifest.
- Results: two history exports, download export, file size/hash comparison and any warnings.
- Finding: four sentences with supporting sources, limits and the next evidence check.
Completing the checklist records your progress; it does not validate the evidence. An explained gap is a useful finding.
Optional extensions¶
The core case is complete. Try one extension in a fresh attempt when you want a new evidence question.
Optional extension: another browser or file-system evidence
Repeat a known action in a fresh Firefox profile. Record its actual profile path from about:support, acquire it consistently, and compare the observations with the Chrome attempt. ChromeHistoryView and MZHistoryView can cross-check supported parsing; a second parser of the same data is not independent corroboration.
For a file-system question, acquire and parse the training volume's MFT/journal using the existing SUSA guides. Match file identity, parent and volume when checking a temporary-to-final rename. Records may not survive. File presence or a rename does not prove execution; process evidence is a separate next check.
Optional extension: searches, bookmarks and cookie metadata
In a fresh, documented attempt, save a bookmark and use the helper's local query and Set benign session cookie links. The helper sets session and one-hour marker cookies; they are not logins. Close/acquire consistently and inspect what actually survives. Closing can discard session data.
Use compatible history/bookmark and cookie viewers. A custom local query is not expected to be recognised as a public engine by MyLastSearch. If approved internet access is available, a harmless query on a documented engine can provide a separate known-action check. Record actual URLs and observations; no result is promised.
Keep cookie lifetime fields separate from visit times and avoid replay/delete/restore operations. A saved link is not a visit, and a cookie does not prove successful authentication. Use history and searches and cookie guidance.
Optional extension: recover a cached image
Use a fresh helper run with an approved existing JPEG and a new case folder:
py -3 C:\Tools\BrowserTraining\serve_lab.py --root E:\Cases\BrowserTraining-Images\site --image E:\TrainingInputs\approved-photo.jpg
Record the actual source hash/size, visit the local page with its embedded photo, stop the helper and acquire the actual profile/cache consistently. Preserve the URL and request context. Cacheable headers do not guarantee retention.
Use a compatible cache viewer to save its listing, then recover selected bytes separately and retain their entry-to-URL-to-output mapping. Compare actual source and output hashes. Review genuine EXIF only if present; do not create camera metadata or invent a recovered image. EXIF dates and export-created file times are not browser visit times.
An unsupported, evicted, partial or no-EXIF result can be valid. Keep the discrepancy and next check rather than forcing a success. See Cache & Recovered Images. No JPEG evidence set is supplied.
Tool controls checked against the official BrowsingHistoryView guide and BrowserDownloadsView guide. Database-copy precautions follow SQLite guidance. Record the versions and behaviour you actually observe.