OWASP DockSec v2026.7.4 and v2026.7.5 Released

DockSec is an OWASP Lab Project that bridges the gap between complex Docker security scan results and actionable fixes.

It combines industry-standard scanners (Trivy, Hadolint, Docker Scout) with multi-LLM support (OpenAI, Anthropic Claude, Google Gemini, or local Ollama) to deliver plain-English vulnerability explanations and line-specific Dockerfile remediation guidance.

Two releases landed today: v2026.7.4 and a quick follow-up fix, v2026.7.5.

v2026.7.4:

## What's Changed

- Fix: HTML report drops all AI findings by @advaitpatel (#139)
- docs: document report formats and CSV header behavior by @ShrijalDubey (#136)
- fix(cli): add verbose log flag by @ahfoysal (#151)
- feat(cli): add --log-file flag by @khldLabs (#152)
- Privacy redaction, waiver file, cache correctness, and slim core install by @advaitpatel (#156)

## New Contributors

- @ShrijalDubey (first contribution in #136)
- @ahfoysal (first contribution in #151)
- @khldLabs (first contribution in #152)

**Full Changelog**: https://github.com/OWASP/DockSec/compare/v2026.7.3...v2026.7.4

Beyond the HTML report and CLI fixes, this release quietly adds a privacy redaction mode and a waiver file for suppressing accepted findings, plus a slimmer core install for setups that don’t need every scanner backend.

v2026.7.5 followed a few hours later as a small hotfix: it removes the deprecated temperature parameter for newer Anthropic Claude models (like Claude Sonnet 5), which had been causing 400 errors - the parameter is now only sent for Claude 2 and 3 generations.

**Full Changelog**: https://github.com/OWASP/DockSec/compare/v2026.7.4...v2026.7.5

Full release notes: v2026.7.4 / v2026.7.5

Happy scanning! :-)

• News
• Tools
• Ai