SLSA
Supply chain Levels for Software Artifacts · Supply-chain Levels for Software Artifacts
A framework of graded build-integrity levels that shows software was built from the source it claims and not tampered with on the way to you.
In our briefings
The claim each of the newest briefings on SLSA would have you leave with.
Ardebili's report keeps a valid signature apart from a known signer. In his words, "A signature alone proves possession of a key." Since v0.105, verification also reports an identity_trust level: self-asserted, local-ca-pinned, ca-anchored or none. An Evidence Bundle carries its own signing key, so a pass on its own proves internal consistency, not identity, unless the auditor pins the expected signer with --public-key. For two-person approval, each signer is bound to an identity the operator vouches for, through a trust list or a pinned CA, a binding he calls "exactly as strong as the operator’s trust list or pinned CA". The check does not yet include the signer's login (OIDC in server mode, or Sigstore). Binding signers to those identities is on his roadmap. Identity elsewhere is early too. Today's trust primitives include did:key identifiers with Verifiable Credentials and delegation ("acted-as") chains, all labelled experimental; workload identity through SPIFFE/SPIRE, and DID/VC identity beyond that did:key slice, sit on the longer horizon. For people, the server mode's OIDC, RBAC, SAML and SCIM support is labelled experimental, and nova login is demo-only.
NovaFabric seals AI agent runs as signed evidence, and its author says a signature proves a key, not a personBI-FIND-2026-1820
RFC 9943 says plainly what a SCITT transparency log does not prove: "registering a Statement only proves it was produced by an Issuer." Issuers can refuse to register statements or submit only some of them, and unless the service's registration policy says so, the order in the log need not match the order of issuance. Against withholding, the RFC puts the burden on whoever reads the log: it calls it important that relying parties not accept a statement they cannot find a receipt for from a service they trust. That sits uneasily with the RFC's first use case, which promises consumers a way to "confirm that sources provide a complete history of statements related to a given component", and with its statement that anyone can review the complete list of statements each issuer registered. We read that completeness as covering what reached the log, not what an issuer kept back. The RFC never mentions AI agents. We read these limits as carrying over to any agent activity log built on it, where an agent that submits its own records could leave some out.
RFC 9943 makes a log prove who signed a statement, not that it is true or that nothing was left outBI-FIND-2026-1407
The paper's central claim is that every agent identity mechanism it surveyed proves which container is acting and nothing about what the agent intends. A trusted execution environment, a zero-knowledge proof or a signed credential can all verify while the model inside has been hijacked by a prompt injection, so the authors propose treating identity as a continuously updated confidence that joins what an agent declared with what it was observed doing. They argue it from a worked example, not a measurement; the TU Berlin prototype of agents with DIDs supplies one, a handshake whose credentials checked out while both agents agreed to skip a step their instructions required.
A survey rates only workload identity ready for agents, and argues no standard can tell what an agent intendsBI-FIND-2026-0644
Cited in 5 briefings
Every briefing that discusses this term, newest first.
- NovaFabric seals AI agent runs as signed evidence, and its author says a signature proves a key, not a person
Paper · · arXiv · Mohsen Seyedkazemi Ardebili · Oct 25, 2026
- RFC 9943 makes a log prove who signed a statement, not that it is true or that nothing was left out
Article · · IETF (RFC Editor) · Henk Birkholz et al. · Jun 30, 2026
- A survey rates only workload identity ready for agents, and argues no standard can tell what an agent intends
Paper · · arXiv · Takumi Otsuka et al. · Apr 25, 2026
- AAGATE maps NIST's AI risk framework onto Kubernetes, and its paper never times the 'millisecond kill switch'
Paper · · arXiv · Ken Huang et al. · Oct 29, 2025
- An early MCP security checklist was never tested, and the protocol has since adopted only part of it
Paper · · arXiv · Vineeth Sai Narajala et al. · Apr 11, 2025
Related terms
What appears alongside SLSA in the same briefings.