Recon creates a lot of data very quickly: subdomain lists, live hosts, screenshots, URLs, JavaScript files, notes and half-finished ideas. Without a system, it ends up scattered across your terminal history and a dozen text files, and you lose the details that lead to bugs. Good bug bounty recon notes fix that. In this post I’ll show a simple folder structure, a note template and a few habits that keep your recon organised from the first session to the final report.
Why bug bounty recon notes matter
Most valid findings don’t come from one command. They come from connecting details: an old endpoint from the archives, a parameter seen in a JavaScript file, a staging host found weeks ago. If those details are not written down, you can’t connect them.
Notes also save time when you come back to a program after a break, help you spot new assets by comparing old and new results, and make report writing much faster because the evidence is already collected.
A simple folder structure
Keep one folder per program and one subfolder per date or per run. Here is a structure that works well:
programs/
example/
scope.md # in-scope, out-of-scope, rules, rate limit, headers
notes.md # your running notes and ideas
findings/ # one file per possible bug, with evidence
recon/
2026-11-20/
subs.txt
live.txt
urls.txt
js-files.txt
2026-12-04/
...The scope.md file comes straight from your reading of the program page. If you haven’t done that yet, follow how to read a program’s scope and rules first.
Keep raw data and notes separate
Tool output is raw data. Your thinking is notes. Keep them apart:
- Raw data goes into the dated recon folders, exactly as the tools produced it. Don’t edit it by hand.
- Notes go into
notes.md: what you looked at, what seemed odd, what you want to test, and what you already ruled out.
This way you can always re-run or compare the data, and your notes stay readable.
A note template that works
For each interesting asset, add a short block to notes.md:
## app.example.com
- First seen: 2026-11-20 (subfinder)
- Tech: React front end, JSON API at /api/v2/
- Auth: email + password, SSO option
- Interesting: /api/v2/export (seen in main.js), old /v1/ paths in archives
- Tested: role changes between two test accounts - nothing yet
- Ideas: check export with the second account; check v1 paths
- Status: in progressThe “Tested” and “Ideas” lines are the most valuable. They stop you from repeating work and remind you where to continue.
Compare runs to find new assets
New assets are often the least tested. If you keep dated results, finding them is easy:
# subdomains in the new run that were not in the old one
sort -u recon/2026-11-20/subs.txt > old.txt
sort -u recon/2026-12-04/subs.txt > new.txt
comm -13 old.txt new.txt > new-subs.txtThe same works for URLs and JavaScript files. The small anew tool by tomnomnom is also handy: it appends only new lines to a file and prints them, so you can see what changed in one step.
subfinder -d example.com -silent | anew all-subs.txtPair this with your subdomain enumeration routine and the archived URLs from Wayback Machine recon to keep your asset list current.
Tools for notes
The tool matters less than the habit. Popular choices include:
- Plain Markdown files in a folder, which work everywhere and are easy to search with
grep. - Obsidian, which adds links between notes and a graph view, useful for connecting assets and ideas.
- Notion or similar apps, if you like databases and tables.
- Mind maps, for a visual overview of a large scope.
Whatever you choose, keep everything private. Recon data about a program, and anything about unfixed bugs, must never be stored in a public repository or shared outside the platform. If you sync notes, use a private, encrypted option.
Recording evidence for reports
Each possible finding gets its own file in findings/. Record:
- The exact URL and request (copy it from Burp Suite or your browser’s developer tools).
- The response that shows the problem.
- The accounts used (always your own test accounts).
- The steps to reproduce, written as you go.
- Screenshots or a short video, with timestamps.
- Your thoughts on impact.
When it’s time to report, most of the work is already done.
Common mistakes with recon notes
- Not writing down what you tested. You will test the same thing twice.
- Overwriting old results. Without old data you can’t spot new assets.
- Mixing programs. One folder per program avoids testing the wrong target.
- Saving secrets or personal data. If you see sensitive data, note what and where, not the data itself.
- Making the system too complex. A simple system you use every day beats a perfect one you abandon.
Conclusion
Bug bounty recon notes turn scattered tool output into a map you can actually use. Keep one folder per program, separate raw data from notes, date every run, write down what you tested, and record evidence as you go. It takes a few minutes per session and pays off every time you return to a target.
If you want to automate the data-collection side so you can spend more time thinking and testing, have a look at the Small Scope Reconnaissance Automation Script, or learn the full workflow with me in the live bug bounty classes.
