Quantum Ventura Request a demo
Source code analysis

CNRT CodeLens

Most scanners hand back a list too long to work through and too noisy to trust. CodeLens is built around the two things that actually make a finding useful: checking its own work, and stopping short of changing your code without being asked.

VerifiedA second model checks the first
ResumableA crash costs one stage, not the scan
Opt-inIt never edits without being asked
CNRT CodeLens · auth/session.c
READY
auth/ session.c token.cnet/ parse.c frame.c
208 size_t need = hdr->len + sizeof(tag_t);
209 if (need > cap) {
210 cap = need;
211 }
212 
213 memcpy(buf, hdr->body, hdr->len);
214 buf[hdr->len] = 0;
215 
216 return finalize(buf, need);
217}
Reachable

Unchecked bounds on copy path

Length comes from the caller and is used to write before it is compared against the destination size.

213line Highseverity 2call sites
Nine stages, resumable
1Index repositoryqueued
2Resolve dependenciesqueued
3Build call graphqueued
4Multi-lens detectionqueued
5Cluster candidatesqueued
6Verify findingsqueued
7Rank by reachqueued
8Draft remediationqueued
9Export reportqueued
A crash costs one stage, not the scan. An interrupted run resumes from its last checkpoint rather than starting again.
auth/session.c:213

Proposed change, for your review

212
213 - memcpy(buf, hdr->body, hdr->len);
213 + if (hdr->len >= cap) return E_RANGE;
214 + memcpy(buf, hdr->body, hdr->len);
ApplySkip Export as MarkdownExport as PDF
Detection stops here. Nothing is written to your repository unless you choose it, finding by finding.
Verification

A second model checks what the first one found

Every finding survives a second opinion.

Detection and verification are separate stages by design. The first pass reads the repository through several different lenses; a later stage takes each candidate finding and asks a different model whether it actually holds. What reaches you has survived being argued with, which is what makes a short list worth reading.

CNRT CodeLens · Verification
ReviewFindingsHistory
STAGE 6
auth/session.c:213

Unchecked bounds on copy path

FIRST PASS CANDIDATE

Length reaches a write before it is compared against the destination size.

SECOND MODEL CONFIRMED

Reachable from two call sites that accept caller-controlled input. Holds.

Rejected at this stage
util/log.c:33 Format string is compile-time constant DROPPED
net/frame.c:88 Guarded by a check the first pass missed DROPPED
Recovery

A crash costs you one stage, not the whole scan

An interrupted scan resumes from its checkpoint.

A full pass over a large repository is long-running work, and long-running work fails. CodeLens checkpoints after each stage, so an interrupted scan picks up where it stopped rather than starting from nothing. On a large codebase that is the difference between a re-run and a lost afternoon.

CNRT CodeLens · Run 1184
ReviewRunHistory
RESUMED
1Index repository00:41
2Resolve dependencies01:12
3Build call graph02:38
4Multi-lens detection08:56
5Cluster candidates01:04
6Verify findingsRESUMED HERE
7Rank by reachqueued
8Draft remediationqueued
9Export reportqueued
The previous attempt stopped partway through verification. This run restarted from that checkpoint, not from the beginning.
Control

It does not touch your code unless you ask

It reports, and waits.

Detection runs first and stops. Remediation is a separate, deliberate step you choose to take, per finding. Plenty of tools will open pull requests against your repository on their own; this one reports, and waits. Findings export as Markdown or PDF so they can go into a review or a report as they are.

CNRT CodeLens · Remediation
ReviewFixesHistory
NOT APPLIED
auth/session.c:213

Proposed change, for your review

211 }
212
213 - memcpy(buf, hdr->body, hdr->len);
213 + if (hdr->len >= cap) return E_RANGE;
214 + memcpy(buf, hdr->body, hdr->len);
215 buf[hdr->len] = 0;
ApplySkip Export as MarkdownExport as PDF
Detection stops here. Nothing is written to your repository unless you choose it, finding by finding.
Where it came from

Built at Quantum Ventura.

Built at Quantum Ventura and sold on its own. It reads source code rather than network traffic, so it stands as its own product, and it reports into the CyberNeuro-RT dashboard for teams already running it.

PositionStandalone product, integrated with the CNRT dashboard
VerificationSeparate model pass over candidate findings
RecoveryCheckpoint and resume between stages
OutputMarkdown and PDF
InterfaceMultilingual, English and Japanese in place today

Tell us the program
and the problem.

We reply from San Jose, usually within two working days.

Request a Demo