Included in the prototype
Six configured Python AST screening heuristics, bounded Git inspection, source identity and integrity evidence, internal review drafts.
READ-ONLY · EVIDENCE-BOUND · HUMAN-REVIEWED
Find potential Python code risks.
Keep the evidence.
A focused, source-bound audit workflow for teams that want concrete review leads—not vague AI security guarantees.
6 configured AST risk checksPinned Git source identityReproducible evidence digests
WHY IT EXISTS
Less guesswork.
More traceability.
Origin identifies candidate code issues with deterministic checks, then binds them to source evidence for independent human review.
01Bounded source inspectionOnly an explicitly permitted, pinned Git revision can be considered for a future customer workflow.
02Evidence-linked leadsFindings retain rule, line and digest information in the internal audit process.
03Human verificationA detection is a review lead, never proof of exploitability or a security certificate.
SYNTHETIC EXAMPLE · NOT A CUSTOMER AUDIT
What a review lead looks like.
DEMO / P001-SAMPLE-001REVIEW REQUIRED
Potential unsafe dynamic evaluation
Risk candidate: dynamic code execution
A deliberately constructed training example contains a call pattern that merits manual investigation. Its presence alone does not prove a vulnerability.
SYNTHETIC
Illustrative rule lead
NOT VERIFIED
CLEAR BOUNDARIES
What this is—and isn't.
Not a security certification
No promise of comprehensive coverage, exploit verification, zero false positives, or an automatically delivered customer audit.
Private code stays private
This website never accepts code uploads. Private-repository access and reports require independently verified consent and Core approval.
PLAN A FUTURE REVIEW
Make a scope draft locally.
This form prepares a downloadable, non-personal scope note entirely in your browser. It does not send an inquiry, transmit files or start a purchase.