You press PrtScn, drag a box, slap an arrow near the broken thing, and upload to prnt.sc. Years of Lightshot have wired that sequence into your hands. Most of it stops helping the moment you move to a browser capture tool where the comment and the crop carry the meaning, not a hand-drawn arrow pointing at a region. Here are the specific reflexes worth dropping, and what to reach for instead.
Drawing an arrow to say "this one" when a pin does the job
In Lightshot the arrow is the whole vocabulary. You point it at the misaligned button and hope the receiver reads where it lands. It is tempting because it is fast and it is the only pointer you have.
A browser tool gives you numbered pins. Drop a pin on the exact spot, and the number ties the mark to the comment you write below it. Pin 1 is the button, pin 2 is the label overlapping it, and your note reads "Pin 2 sits on top of pin 1 at 390px wide." The arrow guesses. The pin is unambiguous, and when the review lists several problems on one screen, numbered pins make the developer read them in order. Keep arrows for direction (this slides off that edge), not for identity.
Uploading the raw shot to prnt.sc and pasting a bare link
The prnt.sc habit is to grab, upload, copy the short link, and drop it in a chat with "see here." The image carries no comment, no context, and no way for anyone to reply on it. It is tempting because it is one keystroke and the link is short.
A published review gives you a /r/<slug> link where every screenshot is an item with your comment attached, and a reader-only /s/<slug> page for people who just need to read it. Anyone with the link can comment on a specific item, and you can mark each comment resolved. That is the difference between "here is a picture" and feedback someone can act on and close out. If you came from Lightshot specifically for the quick-share speed, a browser capture tool that keeps the share fast gives you the link without the dead-end image.
Taking one screenshot per bug and losing the thread
Lightshot is one shot at a time. Five bugs means five uploads, five links, and five messages, and the receiver has to reassemble which link goes with which complaint. You do it anyway because the tool has no concept of a set.
In a browser tool each capture becomes an item inside one review. Walk the page, capture the header, capture the broken form, capture the footer, and they stack into a single ordered list under one link. The receiver opens one thing and reads top to bottom. If you are reporting several problems from a staging build, one review covers the fastest path from bug found to bug filed without juggling a folder of PNGs.
Typing a terse label when you could just say it
On a Lightshot shot you get a tiny text box, so you type "fix this" and move on. Short labels are all the canvas affords, and they strip out everything the receiver needs: what you expected, what happened, where.
A browser tool keeps the short single-line text labels on the image for callouts, but the real comment goes in the item, and you can dictate it with the browser's built-in speech recognition in Chrome or Edge. Talk through what you see at the speed you think, and you end up with a full comment instead of two cryptic words. Firefox users type, since dictation is not supported there. Either way the text label stays a label, and the explanation lives where it has room.
Annotating the live page instead of a frozen still
Some reviewers try to mark up the actual page in the browser, then screenshot the result, so the annotation and the page can drift apart. A browser capture tool grabs the current frame as a still image, and your pins, arrows, boxes, and circles are baked into that still on export. What the receiver opens is exactly what you marked, frozen in time, even after the live site changes. That is why a cropped still beats an annotated live page: the evidence cannot shift under you.
Cropping to hide clutter instead of cropping to point
In Lightshot you crop to trim the taskbar out of frame. The crop is cosmetic. In a browser tool the crop is part of the message: crop down to the broken component and the receiver's eye lands on it before they read a word. A full-page shot with one arrow buried in it makes them hunt. A tight crop plus a pin makes the problem the only thing on screen.
Where the habit swap pays off
None of this needs an install, an extension, or a password. You open a tab, capture or paste an image, pin and comment, and start a new review that publishes to a link anyone can read. It is free to try with unlimited reviews; after 7 days a review goes read-only for you while the shared link keeps working, and Pro at $5 a month keeps everything editable with no cutoff. The output also exports to PDF or Word for people, and to clean markdown at /r/<slug>/markdown if the receiver is an agent. Pick the pin, write the comment, send one link. The Lightshot reflexes you keep are the ones that still point at something real.