Back to inline

Advisories judged against your code

A dependency scanner tells you a package is vulnerable. inline asks the question that matters next: is the vulnerable path reachable from anything you ship? Three tiers of assessment, cheapest first, so triage does not cost what an investigation does.

Confirmed,
or not affected

Every alert is matched to your lockfiles, then assessed against the code that imports it. What comes back confirmed gets a remediation pull request: patched, the suite re-run in a sandbox, and the diff ready to read.

Alertsfeed live

Advisories matched to your lockfiles, then assessed for reachability in your code.

Search package or advisory…
OpenConfirmedAll
PackageSeverityStatusAssess cost
fast-jwt@3.3.1npmCVE-2026-21841
Signature bypass on ES256 tokens
acme/payments-api
criticalconfirmed$0.041
Newtonsoft.Json@12.0.3nugetGHSA-5crp
Deeply nested payloads exhaust the stack
acme/ledger
highconfirmed$0.038
postcss@8.4.31npmCVE-2026-11097
Parser confusion in source maps
acme/web
mediumnot affected$0.009
tar-fs@2.1.1npmGHSA-pq67
Path traversal when extracting archives
acme/mobile
highopen$0.012
urllib3@2.0.6pypiCVE-2026-30820
Redirect leaks the Authorization header
acme/etl
mediumdismissed$0.007
Remediation PR openedverified

feature/inline-fix-cve-2026-21841

Bump fast-jwt to 4.0.2 and pin the ES256 curve

  • Reachable from POST /v1/charges
  • Patch applied and the suite re-run in a sandbox
  • No behavior change outside token verification
+18−63 files

And a red team
that proves it

An opt-in offensive run maps your attack surface from the code and the graph, fans out one agent per attack class, then validates the promising leads by building a non-destructive proof in a sandbox. Only what it proved is reported.

Pentest findingsacme/payments-api · validated
Validated

4

Candidates

23

Run cost

$3.14

IDOR on GET /v1/charges/:id
tenant scope missing · PoC proven in sandbox
critical
Prompt injection via the PR title
reaches the review tool loop
high
Rate limiter keyed on a spoofable header
X-Forwarded-For trusted directly
high
Verbose stack traces in 500 responses
leaks file paths and query text
medium

Included

Threat feedadvisories matched the hour they publish
Reachabilitythree tiers: triage, deep pass, escalate
Remediation PRson a branch prefix you configure
Prompt securityinjection and tool-abuse classes, run against your own prompts

Put it on one repository

Connect GitHub, enable a repo, and the next pull request opened gets a review.