Signing Is Not Authorization: Claim-Level Negative Vectors for Agent-Payment Receipts

A valid signature establishes that particular content was authenticated under a particular key. By itself it does not establish that the key belonged to the claimed authority, that the attested action occurred, or that the action was authorized under the parameters actually used. For agent-payment receipts these are separable properties. This note records a dated position and a reproducible characterization: the separation is executable — receipt-bound negative vectors reject authenticated-but-claim-invalid payment receipts, each for a distinct semantic reason, while envelope signature verification succeeds. The contribution is not the established observation that signatures do not prove authorization; it is the executable four-property receipt decomposition applied specifically to agent-payment evidence, extending the general method deposited in concept DOI 10.5281/zenodo.21418701. It does not claim to originate the adversarial-conformance layer, which has public prior work. Evidence: Ed25519-signed receipt-claim conformance vectors RCL-001..011 (9 rejections + 2 positive controls), deterministic and independently recomputable.