Kusari secures the software supply chain for regulated companies with complex technology stacks. It helps security and engineering teams in financial services, insurance and regulated SaaS see every dependency they ship, decide which vulnerabilities actually matter and fix them before code reaches production.
The problem it targets is noise: teams drown in scanner findings yet cannot show a regulator which risks are real and who is fixing them.
At A Glance
| Vendor | Kusari |
|---|---|
| Category | Software supply chain security |
| Agentic AI | Policy for agent-written code, MCP server for coding agents |
| Market classification | DevSecOps |
What Kusari Does
Most organisations already run scanners, registries and pipelines, yet still cannot answer simple questions such as which applications use a vulnerable package or who owns the fix. Kusari builds one verified graph of every dependency, assembled from source, and sits as an intelligence layer above the existing tools rather than replacing them.
Risk is ranked from that graph across the estate and checked again at each code change, and the evidence maps to the frameworks regulated buyers answer to, including DORA, NYDFS Part 500, PCI DSS 4.x, FFIEC, HIPAA, the EU CRA, FedRAMP, SOC 2 and FDA 524B.
Products
Kusari Platform
The Platform is the command centre for supply chain risk across an organisation’s repositories. It helps an application security team cut a long vulnerability backlog down to the items an attacker could reach, and then get those items fixed by the right team.
| Reachability | Ranks what is exploitable, not just what is present |
|---|---|
| Blast Radius | Traces one package to every application, service and owning team |
| AutoFix | Drafts the fix and proves it in the customer’s CI, leaving the merge to the team |
| Security Assistant | Answers estate wide questions from the verified graph |
| Policy and attestation | Rules such as blocking unsigned code and unattested artifacts, based on signed commits and SLSA provenance |
| VEX and scoring | Findings carry “VEX with reasoning” and a Kusari Score |
Kusari Inspector
Inspector gives developers a security review on every pull request, checking the code and its dependencies several layers deep and returning a merge decision together with a fix. It suits teams that want problems caught in review instead of after release.
- Covers known vulnerabilities, transitive dependencies, secrets, typosquatting, licences, unmaintained components, code weaknesses, and pipeline or container configuration
- Runs established open source analysers for each ecosystem and drops results that cannot be reached
- Applies organisation level policy, with a
kusari.yamlfile at the repository root taking precedence and exceptions needing admin approval - Supports Go, Node.js, Python, Java, .NET, Ruby, Rust, Terraform, OpenTofu, Docker and Helm
Waybill
Waybill is an open source SBOM generator built and maintained by co-founder Michael Lieberman. Because the Platform takes SBOMs as its common input, Waybill extends it to new build systems.
How It Works
| Component | What it does |
|---|---|
| Ingestion | Every source of evidence enters as an SBOM or attestation, in CycloneDX, SPDX, OpenVEX or SLSA provenance format. |
| Dependency graph | The Platform builds the verified graph from source and works across GitHub, GitLab and Bitbucket. Findings can be enriched with Black Duck results and routed to Jira for the owning team. |
| Pull request review | Inspector runs as a GitHub App that posts its verdict as a pull request comment, as a GitHub Action, on GitLab merge requests, through CI/CD templates and through the Kusari CLI. |
| CLI | The command-line tool connects to both Inspector and the Platform from any environment. |
| Limits | Inspector handles up to 2,000 dependency changes per pull request, examining up to 1,000 high priority changes in depth. |
Agentic AI Capabilities
Policy For Agent-Written Code
Inspector policy applies “the same way whether a person or a coding agent wrote it”. That matters for typosquatting in particular: Inspector flags look-alike packages even when a coding agent suggested them, closing a route by which AI assistants can pull malicious dependencies into a project.
MCP Server For Coding Agents
An MCP server lets a coding agent call Inspector’s security check while it is still writing code, so findings return to the IDE or agent before a pull request exists. The CLI and IDE integrations offer the same checks for local use.
Company Background
| Founders | Tim Miller (co-founder and CEO), Michael Lieberman (co-founder and CTO), Parth Patel (co-founder and CPO) and Brian Miller (co-founder and CRO). |
|---|---|
| Open source | The founding team created GUAC, built with Google and now an OpenSSF Incubating Project, and contributes to SLSA. |
| Funding | $8M raised from J2 Ventures, Glasswing Ventures and Unusual Ventures. |
Note: This overview is compiled from Kusari’s publicly available information as of 3 November 2026. It is provided for information only. NHI Mgmt Group does not endorse any vendor, and nothing on this page represents the opinions of NHI Mgmt Group.