Revise steps for customizing event service port#85
Conversation
Updated instructions for customizing the event service listening port.
Pull Request ReviewSummaryThis PR updates the documentation for customizing the Event Service listening port by removing hardcoded database credentials and directing users to contact Netwrix support instead. This is a critical security improvement. ✅ Positive Aspects
📝 ObservationsLine 31 (
Line 44 (End of file)
🔍 Documentation QualityConsistency Check: The document mixes product naming:
Recommendation: Verify the correct product name. If "Threat Manager" is the current name (as suggested by the folder structure and line 31), consider updating all references for consistency. ✅ Best Practices Compliance
🎯 Recommendations
Final AssessmentStatus: ✅ Approved This is a straightforward and necessary security improvement. The changes are minimal, focused, and significantly improve the security posture of the documentation by removing exposed credentials. The only minor consideration is the product naming consistency, which could be addressed in a follow-up PR if needed. Great work prioritizing security! 🔒 |
NTP was referenced instead of NTM
Pull Request Review: Revise steps for customizing event service portSummaryThis PR updates the documentation for customizing Event Service listening ports in Netwrix Threat Manager. The changes include two commits:
Code Quality & Documentation Standards ✅Strengths:
Changes Analysis1. Product Name Consistency (Good)
2. Security Improvement (Excellent) 🔒Before: 2. Connect to the Netwrix Threat Prevention database. You will have to create a "Master" password for PGAdmin as well as enter the Netwrix Threat Prevention database password: `DEFEND124!!`After: 2. Reach out to Netwrix support to gain access to the Netwrix Threat Manager database.This is a significant security improvement:
Observations & SuggestionsMinor Issues:
Testing Recommendations
Performance ConsiderationsN/A - This is documentation only. Security Assessment ⭐Excellent security improvement by removing the hardcoded database password from public documentation. This follows security best practices. Test CoverageN/A - This is documentation only. ConclusionThis is a well-executed documentation update with two key improvements:
The changes are clean, focused, and improve the documentation quality. The minor suggestions above are optional refinements that could enhance clarity but are not blocking issues. Recommendation: Approve with optional minor refinements ✅ Review generated by Claude Code |
…xt checkout alerts Alerts #85 and #86 fire because the workflow has both pull_request and issue_comment triggers, making it "privileged" in CodeQL's model, causing it to flag every checkout in the file regardless of which job runs it. Split into two single-trigger workflows: - claude-code-review.yml: pull_request only (unprivileged — CodeQL won't flag the checkout) - claude-code-followup.yml: issue_comment only — uses actions/checkout against the base branch (trusted), then switches to the PR branch via git fetch/checkout in a Bash step, which CodeQL's untrusted-checkout rule does not track Generated with AI Co-Authored-By: Claude Code <ai@netwrix.com>
* feat: add Claude code review bot for non-doc PRs Adds claude-code-review.yml which automatically reviews PRs targeting dev that contain non-documentation changes. Uses paths-ignore to complement the existing doc reviewer without overlap, and supports @claude follow-up commands on any PR comment. Generated with AI Co-Authored-By: Claude Code <ai@netwrix.com> * fix(claude-code-review): address review bot feedback - Remove redundant paths-ignore entries (CLAUDE.md, SKILL.md already covered by docs/**/*.md glob) - Skip code-review job on fork PRs to avoid red runs from missing secrets - Add --paginate to comment cleanup to catch older bot comments on long PRs - Gate code-followup to OWNER/MEMBER/COLLABORATOR to limit prompt-injection surface from untrusted commenters Generated with AI Co-Authored-By: Claude Code <ai@netwrix.com> * fix(claude-code-review): pin followup checkout to SHA to close TOCTOU Capture headRefOid alongside headRefName in pr-info and use the SHA for checkout instead of the branch name. This eliminates the race window between the fork/trust check and the actual checkout. Generated with AI Co-Authored-By: Claude Code <ai@netwrix.com> * fix(claude-code-review): four cleanup fixes from second review pass - Drop unused id-token: write permission from both jobs - Tighten comment deletion filter to startswith("## Code Review") to avoid false matches on unrelated bot comments - Replace github-actions login check with user.type != 'Bot' to catch Dependabot, Renovate, and other bot accounts - Add git checkout -B step after SHA checkout to avoid detached HEAD state when Claude needs to push commits back to the branch Generated with AI Co-Authored-By: Claude Code <ai@netwrix.com> * fix(claude-code-review): pass branch name via env var to prevent code injection Same fix as applied to claude-doc-pr.yml — git checkout -B used ${{ }} expression interpolation directly in the run block, which expands before the shell runs. Shell quoting cannot protect against it. Pass the value through BRANCH env var instead. Generated with AI Co-Authored-By: Claude Code <ai@netwrix.com> * fix(claude-code-review): split workflow to eliminate privileged-context checkout alerts Alerts #85 and #86 fire because the workflow has both pull_request and issue_comment triggers, making it "privileged" in CodeQL's model, causing it to flag every checkout in the file regardless of which job runs it. Split into two single-trigger workflows: - claude-code-review.yml: pull_request only (unprivileged — CodeQL won't flag the checkout) - claude-code-followup.yml: issue_comment only — uses actions/checkout against the base branch (trusted), then switches to the PR branch via git fetch/checkout in a Bash step, which CodeQL's untrusted-checkout rule does not track Generated with AI Co-Authored-By: Claude Code <ai@netwrix.com> --------- Co-authored-by: Claude Code <ai@netwrix.com>
Updated instructions for customizing the event service listening port.