fix(compose): surface per-service scan failures in user-facing summary

[myukitty]

Pull Request

Description

Rebased onto current main and rewritten now that #161 has merged.

#161 added failed_services tracking inside ComposeOrchestrator.run_full_scan, which covers the first two acceptance criteria of #131. Nothing reads that field back out, so the third one is still open: a compose run where some services could not be scanned still prints a Quick take and a security score as if every service had been covered. That is the misleading behaviour #131 was filed about.

This PR adds only the missing user-facing half, on top of #161:

1. compose_scanner.py carries total_services alongside failed_services, so the summary has a denominator.
2. cli.py prints one Quick take line — N of M services could not be scanned: <names> — and includes failed_services / total_services in --json output.
3. A service that fails both its Dockerfile scan and its image scan is appended twice by run_full_scan, so names are de-duplicated before counting.

No change to exit codes, scoring, or scan behaviour — this is purely the visibility fix #131 asked for.

Before (current main, 2 of 3 services unscannable):

Quick take
  - 16 security findings (0 critical, 0 high)
  - Run without --scan-only to add AI-powered explanations and fixes

After:

Quick take
  - 16 security findings (0 critical, 0 high)
  - 2 of 3 services could not be scanned: web, db
  - Run without --scan-only to add AI-powered explanations and fixes

The per-service reason strings are available in --json under scan_info.failed_services; I kept them out of the one-line Quick take because they vary per service. Happy to inline the reason if you would rather match the issue's example wording exactly.

Closes #131

Type of Change

• Bug fix

How Has This Been Tested?

• Unit tests
• Manual testing

Added 7 tests: 4 in tests/test_cli.py covering the summary line, the de-duplication, the no-total fallback, and silence when nothing failed; 3 in tests/test_compose_scanner.py covering total_services, populated failed_services, and the unparseable-compose path.

• pytest tests/ — 254 passed, 2 skipped. The 2 failures in tests/test_utils.py on my machine are the optional docksec[ai] extras not being installed, and reproduce unchanged on main.
• ruff check . — all checks passed.
• Ran the real orchestrator and renderer against a 3-service compose file with 2 unresolvable images to produce the output above.

Test Configuration:

• Python version: 3.12.13
• Operating System: Windows 11
• DockSec version: main @ 77c0cd1

Checklist

• Code follows the style guidelines of this project
• Self-review completed
• Hard-to-understand areas are commented
• No new warnings or errors introduced
• Tests added that prove the fix works
• All existing tests pass
• Spelling checked

Related Issues / PRs

• Closes [GOOD FIRST ISSUE] Surface per-service compose scan failures in user-facing output #131
• Builds on fix: surface failed compose service scan information #161

---

By submitting this pull request, I confirm that my contribution is made under the terms of the MIT license.

[advaitpatel]

Hello @myukitty - can you please update your branch and PR and resubmit this PR again?

[myukitty]

Thanks @advaitpatel — updated and rebased onto current main (77c0cd1); the branch is mergeable again.

Since #161 merged in the meantime, I rewrote the PR rather than just resolving the conflict. #161 already tracks failed_services in run_full_scan, so I dropped everything this PR duplicated and kept only the part of #131 that is still unfixed: nothing reads that field back out, so a compose run where some services could not be scanned still prints a Quick take and a score as if every service was covered.

What is left is 4 files, +141/-3:

• total_services carried alongside failed_services so the summary has a denominator
• one Quick take line: 2 of 3 services could not be scanned: web, db
• failed_services / total_services in --json
• de-duplication, because a service that fails both its Dockerfile and image scan is currently appended twice
• 7 tests

No change to exit codes, scoring, or scan behaviour. Full details and before/after output are in the updated description.

[codecov-commenter]

Codecov Report

❌ Patch coverage is 95.65217% with 3 lines in your changes missing coverage. Please review.
✅ Project coverage is 80.51%. Comparing base (a2fe184) to head (e385a53).
⚠️ Report is 27 commits behind head on main.

Files with missing lines	Patch %	Lines
docksec/cli.py	85.00%	3 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #160      +/-   ##
==========================================
+ Coverage   78.95%   80.51%   +1.56%     
==========================================
  Files          26       32       +6     
  Lines        4158     5169    +1011     
==========================================
+ Hits         3283     4162     +879     
- Misses        875     1007     +132     

Flag	Coverage Δ
unittests	80.51% <95.65%> (+1.56%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[advaitpatel]

lgtm