Around 3400 BCE, in the floodplains of southern Iraq, Sumerian scribes cut reeds from the riverbank, trimmed one end flat, and pushed it into damp clay tablets at different angles. The marks came out wedge-shaped, which is where we get the word cuneiform, from the Latin cuneus, “wedge.” The reed itself barely qualified as a tool. A cut length of plant. What did the real work was the clay: wet enough to dent, then left in the sun or baked hard so the dents stayed put. Thousands of those tablets survive. Most of them are receipts. Barley deliveries, sheep counts, who owed what to which temple. Accountancy, basically.
Push the timeline back further and the tools get simpler still. People were scratching figures into cave walls and rock faces by dragging a harder stone across a softer one, flint across sandstone, long before anything we’d call a writing system existed. Whether petroglyphs count as writing is an argument for another post. The mechanics are the same either way. Hard thing, soft thing, mark.
Egyptian scribes around 3100 BCE had a different problem, because papyrus won’t take an impression the way clay will. Their solution was a chewed-up marsh rush. You took a stem of Juncus, bashed or chewed the tip until the fibres frayed out, and used the resulting bristly end to pick up ink made from soot and water. Crude, but it worked on papyrus, and it worked on plastered tomb walls.
Bone showed up everywhere. A leg bone from a sheep, sharpened against a stone, made a perfectly good stylus for pressing into clay or scoring wax. Cheap, sturdy, lying around on the butcher’s floor. The Romans later used metal styluses for their wax tablets, but bone versions turn up in the archaeology too, especially in cheaper contexts.
So the real answer to “what were the first writing instruments” is deflating. A reed. A stone. A bone. A plant stem with the end chewed flat. People wrote with whatever the riverbank handed them.