Documentation
Use Cases
Worked end-to-end examples covering triaging a CVE, gating a dependency, running an incident, hitting a KEV deadline, and reviewing a PR.
Each guide is a real sequence: what to ask, which tools answer, what comes back,
and what you do with it.
The pattern underneath
Every guide follows the same shape, because it is the shape of the work:
- Establish the facts. KEV, EPSS, exploitation maturity and sightings,
rather than what the model recalls.
- Establish the exposure. Does it reach your code, from where, reachable by
whom.
- Decide, and record why. The decision matters less than being able to
defend it in three months.
- Act, then verify. A fix nobody confirmed is a hope.
Tools cover step 1 completely. Steps 2–4 need your repository, which is where
the local prompts and the CLI come in.
Triage a CVE properly
From a bare identifier to a decision you can defend, using real data instead of a severity label.
Gate a new dependency
Check a package before it enters the lockfile, covering vulnerability history, exploitation signals, typosquat risk, maintenance health and a safe version.
Hit a KEV deadline
Find binding remediation deadlines that apply to what you actually run, and turn them into an ordered plan.
Run an incident
A CVE goes hot. Confirm urgency, pull IOCs and detection content, decide on a patch path, and record the position, all in one conversation.
Review a pull request
Security review of a diff, covering new dependencies, changed code, and a gate that can fail the build.
Audit an inherited repo
First contact with a codebase nobody has looked at. What to check, in what order, and how to stop before you drown.