signsan

SignSan is a framework proposal in the field of digital security that aims to combine cryptographic signing with content sanitization to maintain authenticity while preventing unsafe content alterations during processing. The core idea is to sign a canonical representation of content after it has been sanitized according to a defined policy, and to attach a manifest describing the sanitization rules used. A signer uses a digital signature on the canonicalized, sanitized form; a verifier re-applies the same sanitization policy to the received content, reconstructs the canonical form, and checks the signature.

Key components include a signing policy, a sanitizer, a canonicalizer, and a verification library. Sanitization policies may specify allowed transformations, such as removing script tags, normalizing whitespace, escaping or stripping disallowed attributes, and enforcing consistent encoding. The approach emphasizes end-to-end trust by ensuring that what is signed is exactly what is later verified after sanitization steps have been applied.

SignSan is not yet standardized; various projects experiment with its principles in content distribution, document signing, and software provenance. Adoption challenges include defining precise, interoperable policies, managing policy versioning, and handling potential performance overhead. Limitations include potential policy drift, the possibility that content could be augmented after signing in ways the policy does not anticipate, and dependence on trusted policy publishers.

See also: digital signature, content sanitization, canonicalization, verifiable credentials.