Mtictactoesweeper

Niche indie game you just totally haven't heard of, duuuude.

Usage

Clone, then start up a webserver in the repo's root. python3 -m http.server (aka python2 -m SimpleHTTPServer) suffices, for example.

Each player starts with a bit more than half as many flags as there are mines. (5/8 rounded down, to be exact)
On each turn, a player can either reveal or flag/unflag a covered tile.
Like in regular minesweeper, a tile cannot be uncovered if it is flagged.
A flag can only be removed by the player who placed it.
Flags can only be placed on covered tiles that touch either (a) an uncovered tile or (b) a flag of the same color, i.e. placed by the same player.

Unlike in regular minesweeper, the game does not end when all non-mine tiles are uncovered.
Instead, the game can be...

won by placing three or more flags of the same color in a straight line (tic-tac-toe victory)
won by having flagged the most mines correctly once all mines are flagged (minesweeper victory)
lost by uncovering a mine (minesweeper defeat)

Uncovering tiles rashly to stop an opponent from advancing toward tic-tac-toe victory.
Flagging non-mines on purpose (particularly when sandwiched in between two other mines) to preempt the above.
On running out of flags, having to choose tactically which previously-placed flags to take back (which can open up room for tic-tac-toe defeat). Overall, and particularly with low mine-counts, the game plays more like "tactical tic-tac-toe" than like minesweeper. That could be mitigated by bumping up the number of mines... but that also makes it even-more boring, so.

Phaser 3. Rather enjoyable to work with, even if there aren't yet too many non-Phaser-2 resources out there. Would recommend.
Discord for introducing spoiler tags, which prompted people to make minesweeper games and ultimately got this idea into my head