
From Dependency Lists to Attack Paths: Building Blast Radius with FalkorDB
Blast Radius turns npm dependency and vulnerability data into a graph you can explore. See how FalkorDB powers attack-path queries, remediation planning, and what-if simulations—with a local SQLite fallback.
From Dependency Lists to Attack Paths: Building Blast Radius with FalkorDB
Dependency vulnerabilities are often presented as a long list: package name, advisory, severity, and a suggested version. That is useful, but it does not always answer the questions a developer needs to act:
- How does this vulnerable package reach my application?
- Which other packages depend on it?
- Which upgrade could address the most risk?
- What might change if I choose that upgrade?
Blast Radius is a supply-chain security triage workspace built to explore those questions through a dependency graph.
Turning package data into a graph
Blast Radius retrieves npm dependency information from deps.dev and public vulnerability advisories from OSV. It represents the application, package versions, advisories, and candidate fixes as graph entities connected by relationships such as
DEPENDS_ON , HAS_VULN , and RESOLVES .For the included Kitchen Sink demo profile, the graph contains 1,278 packages and 76 advisories. FalkorDB is the graph engine for traversals and read-only Cypher queries. I verified the application reporting FalkorDB as its active engine and exercised all ten graph tools against the local database.
Following risk through dependencies
A vulnerable package is more informative when we can see how it connects to the application. Blast Radius can trace an attack path, explore package blast radius, identify high-impact dependencies, and group packages into risk clusters.
The graph view makes those relationships explorable instead of leaving them buried in separate tables. The analyst workflow can call graph tools and present their results as evidence alongside the UI.
Prioritizing possible fixes
Blast Radius builds a remediation plan that ranks candidate package changes by risk reduction relative to estimated effort. It distinguishes upgrades that fit existing requirements from overrides and potentially breaking changes. Advisories without a suitable upstream fix are reported separately rather than counted as resolved.
A proposed fix can be simulated before it is applied. Simulation is only a preview: it estimates the remaining risk from the selected fixes and does not modify a project’s package files. The app can also generate suggested package overrides and review commands for a developer to inspect.
Keeping local development flexible
FalkorDB is optional for local use. Blast Radius also has a SQLite snapshot store and uses NetworkX for graph analysis, so the basic application can run without Docker. The interface reports which graph engine is active.
The analyst has a separate mode indicator. When a Gemini API key is configured, it can run in live mode; without a key, it uses recorded replay data. Replay is useful for a predictable demo, but it should not be presented as a live model response.
What I learned
The central design lesson was to make the graph do meaningful work. A graph visualization alone is not enough: the data model, queries, remediation logic, and interface need to agree about what a dependency or fix means.
It is also important to distinguish analysis from action. A simulation should not imply that a package has actually been upgraded, and an advisory should not be marked fixed unless the proposed version really resolves it. Clear engine and agent-mode indicators help users understand where results came from.
Blast Radius is a triage aid, not a replacement for reviewing dependency changes, testing an application, or validating security findings. Its purpose is to make dependency relationships and remediation trade-offs easier to inspect before a developer decides what to change.
Enjoyed reading this content? Let the author know!
Your likes, comments, shares, and saves help creators reach more builders.
Loading recommendations
Loading article