Post preview.

Monero's privacy is not one clever trick. It is three, stacked, each plugging a different leak. Knowing which does what is the line between a threat model and cargo-cult OPSEC.

Ring signatures. Your real output is mixed with decoys pulled from the chain. A verifier can confirm one member of the ring signed, but not which one. The sender is hidden inside a crowd. Current ring size is 16.

FCMP++ (testnet live, mainnet pending) swaps the sampled ring for a proof over the entire output set. The crowd stops being 16 and becomes the whole chain. This is the upgrade worth actually watching. Also: not shipped yet, so do not quote it as if it is.

Stealth addresses. The sender derives a one-time destination from the recipient's public key. Nobody on-chain can tie two incoming payments to the same recipient without their private key. Address reuse is not a problem you can have on Monero. The protocol won't let you.

RingCT. Amounts hidden with Pedersen commitments. The network checks that inputs equal outputs plus fees, so nobody can print coins, without ever seeing a figure. The amount is cryptographically absent, not rounded off or starred out.

Stack them and a passive observer gets a set of maybe-senders, a one-time address that has never existed before, and no amount. The three axes of financial surveillance all close at the same time.

The honest limit: ring signatures are probabilistic, not absolute. Timing and exchange-deposit correlation have narrowed real-world sets before. Private Mode applies these properties to the XMR leg. Your source and destination chains are still your problem to keep tidy.