Security Architecture
Autonomous AI agents require defense-in-depth package distribution. Agent SPM enforces cryptographic verification, strict schema typing, and zero-runtime leakage across the entire supply chain.
SHA-256 Cryptographic Signing
Every published skill release is digested and cryptographically signed on publish. When the SPM CLI runs spm install, it verifies the local hash against the immutable registry ledger to prevent tamper or man-in-the-middle substitution.
Strict JSON Schema Validation
Agent procedural tools require deterministic inputs and outputs. SPM enforces runtime schema checking using TypeBox and Zod definitions, preventing prompt injection attacks from manipulating parameter types.
Isolated Private Namespaces
Enterprise skills reside in dedicated private tenant spaces with row-level encryption at rest (AES-256) and in transit (TLS 1.3). Cross-tenant access is physically impossible at the database protocol level.
Automated Vulnerability Scanning
Every package uploaded is analyzed by automated static analyzers checking for dangerous child process invocations, unverified network socket bindings, and known CVEs prior to registry publication.
01Vulnerability Disclosure & Bug Bounty
We welcome vulnerability reports from independent security researchers. If you identify a security flaw or supply-chain vulnerability, report it immediately to security@agent-spm.com. PGP keys are available upon request.
02SOC 2 Type II Report Access
Enterprise customers on our Lifetime Enterprise plan can request our full third-party SOC 2 Type II audit report, penetration testing summary, and ISO 27001 certificate directly via the Compliance Portal.