Source code security scanning
Know what is in the code before you run it.
CodeGuard inspects uploads, repositories, branches and pull requests for malware indicators, vulnerable dependencies, exposed secrets and risky configuration — and explains every finding with evidence, coverage and limitations.
What it scans for
Malware indicators
Obfuscation, remote code execution, credential theft, exfiltration and backdoor patterns — reported as indicators for review, with evidence.
Vulnerable dependencies
Manifests and lockfiles matched against the OSV database. Advisory IDs, affected ranges and published fixed versions only.
Exposed secrets
Provider-specific formats and high-entropy strings. Evidence is always masked — full secret values are never stored or shown.
Configuration risks
Risky install scripts, CI configuration, exposed files and insecure settings that widen your attack surface.
Git history
Commit-by-commit signals: dependency and install-script changes, new URLs, executables and config changes.
Pull requests
Only what a change introduces relative to its target branch, with links to the exact lines in the diff.
How it works
- 1
Bring your code
Upload files, a folder or an archive; paste a GitHub, GitLab, Bitbucket, Dropbox or Google Drive link; or connect a repository.
- 2
Isolated scanning
Workers fetch the source, build an inventory and run rule, secret, dependency and history engines — then score the results.
- 3
Explainable report
Findings classified as Confirmed, Suspicious or Possible false positive, with coverage and limitations listed alongside.
- 4
Chrome extension
Scan the repository, file or pull request you are viewing on GitHub, GitLab or Bitbucket with one click, and open the full report here.
An honest note on what scanning can and cannot do
No automated scanner — including CodeGuard — can guarantee code is 100% free of risk. We believe in total transparency about capabilities and limitations.
What it finds
Known CVEs, exposed secrets, risky install scripts, and malicious indicator patterns with evidence.
Known limitations
Novel zero-days or deep obfuscation can be missed. Some matches require human triage.
Explainable score
Risk score is additive and explainable — not a safety guarantee. Treat results as input to review.
Plans
Scan your first project in minutes
Upload a folder or paste a repository URL. Reports are private to your organization.