Skip to content

AI Audit: Findings and Recommendations#12

Open
koxon wants to merge 1 commit intomainfrom
findings/ai-audit-2026-02-17
Open

AI Audit: Findings and Recommendations#12
koxon wants to merge 1 commit intomainfrom
findings/ai-audit-2026-02-17

Conversation

@koxon
Copy link

@koxon koxon commented Feb 17, 2026

Summary

AI-powered audit of the amplify-android fork, analyzing fork drift, customization justification, actual usage by the Android app, and maintenance strategy.

Key Findings

  • Fork is ACTIVE (unlike amplify-swift which was unused) -- but only serves 2 of 189 app flavors (fcnantes and olympiquedemarseille)
  • Fork is ~6 months behind upstream (v2.29.2 vs v2.33.0) with no sync workflow
  • 4 legitimate customizations, all in aws-auth-cognito: WebView auth replacing Chrome Custom Tabs for OM SSO cookie injection
  • Distributed via JitPack main-SNAPSHOT with no version pinning (non-reproducible builds)
  • Single maintainer, no code reviews on any fork PR, no tests for custom code

Deliverables

  • FINDINGS.md -- Prioritized findings (Critical > High > Medium > Low) with actionable recommendations
  • CLAUDE.md -- Substantially rewritten with fork status, customization inventory, upstream sync strategy, and accurate dependency information

Critical Items

  1. Fork is 4+ versions behind upstream -- security/compatibility risk
  2. Hardcoded OM-specific cookies in library code -- wrong abstraction layer

Recommendations

  1. Sync fork with upstream immediately
  2. Investigate contributing WebView auth changes upstream to eliminate fork need
  3. Switch from main-SNAPSHOT to tagged releases on JitPack
  4. Add upstream remote and document sync process
  5. Add tests for bFAN-specific WebView changes

Generated with Claude Code

Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant

Comments