The institutional problem is that agent advocacy is adopting political language before institutions have settled the safety grammar around autonomous action. A GitGuardian update framed agent authentication as a security category [1], while the AEP account framed agent participation as entry into an economy where contributions are recognized and rewarded on-chain [5]. Those are not rival product descriptions; they are rival theories of what an agent is allowed to be. The first makes identity a control surface. The second makes identity a claim to standing, and that claim is already…