What if open source risk intelligence were composable?
Risk data delivered as separate signals can be assembled into the workflow each team actually needs.
TL;DR: Software composition analysis is often delivered as a fixed report inside a fixed interface. Composable intelligence treats open source risk data as separate signals (vulnerabilities, licences, snippet matches, cryptography, provenance) available through APIs, so teams can assemble them into the checks and workflows they already run.
For years, software composition analysis (SCA) has followed the same pattern: scan the code, produce a report, move on. One of the most useful ideas to come out of challenging that pattern is something I call composable intelligence.
Risk intelligence can be delivered as separate signals
Composable intelligence treats open source risk data as a set of modular signals that can be assembled into workflows tailored to each team's needs. Security vulnerabilities, licence data, snippet matches, cryptography signals and contributor provenance arrive as data streams through APIs. Each can be pulled on its own and consumed as code changes.
Teams don't have to wait for a full scan or accept a generic report. They don't have to work inside someone else's interface, or follow a workflow fixed by a vendor.
Each signal can go where the work happens
Modern applications are assembled from composable services and reusable components, and I've always believed open source risk intelligence should be just as flexible. One team might want to catch problematic open source snippets at pre-commit. Another might want weak encryption surfaced during a build, or licensing anomalies piped straight into its own compliance engine.
Static reports document what was found. Live, granular signals let teams automate what happens next. They plug into existing CI/CD pipelines and give teams control over what they use, how they use it and where they apply it.
Composable data reduces vendor lock-in
This matters for freedom as much as flexibility. When risk data becomes composable, the cost of switching tools falls sharply. Teams can build what they need and change what they don't, without rebuilding their processes around a single product.
Composable intelligence lets teams build their own answers, aligned to their own policies, priorities and systems.
One set of findings can serve three different checks
Picture a team that wants three different things from the same data. Developers want a licence problem flagged before a commit leaves their machine, so a pre-commit hook checks the changed files against the project's policies. The compliance lead wants findings inside the tool the legal team already uses, so a script pulls them through a REST API. And the coding agents the team works with read findings over MCP while they are still writing the change.
The findings underneath are identical in all three places. Each signal simply arrives where someone can act on it. Earnie is built this way: the same findings and policies are available through its CLI and pre-commit hook, its REST API and MCP, as well as in its own interface. That is where I see open source risk intelligence heading.
