[AAASM-3946] 📝 (docs): Test AI onboarding flow with real usage scenarios

[Chisanan232]

Jira Ticket

• AAASM-3946

Summary

• Dogfoods the org AI onboarding scaffold built by AAASM-3939–3945 by actually reading every produced artifact cold and exercising it through 4 contributor personas (first-time contributor, returning maintainer, workspace setup, task preparation), plus an explicit Claude Code vs. Codex coverage comparison. Depends on [AAASM-3939] 📝 (workspace): Add org-level AI workspace layout doc #25-[AAASM-3945] 📝 (docs): Apply AI onboarding scaffold to selected repos #31 (full chain).

Changes

• Add docs/onboarding-poc/AAASM-3946-dogfooding-notes.md: 4 scenario write-ups, the Claude Code/Codex comparison, and a numbered list of 12 concrete improvement items for AAASM-3947 to triage.

Validation

• Content reviewed for accuracy
• Links verified
• No unintended scope changes

Checklist

• PR title follows: [<ticket ID>] <GitEmoji> (<change scope category>): <key point as summary>
• PR body follows this template

[Chisanan232]

Claude Code review — AAASM-3946

CI: Red, same org-wide Actions billing-block (see #25) — not caused by this change.

Scope vs ticket: Matches AAASM-3946's AC well beyond the minimum bar — all 4 personas (first-time contributor, returning maintainer, workspace setup, task preparation) were actually exercised against the real scripts/docs (not just described), and the Claude Code vs Codex asymmetry is documented concretely.

Standout: this ticket caught a real, verified defect (the WORKSPACE.md-vs-validate-ai-workspace.sh per-file override mismatch — see my comment on #30) by actually building a fake override and running the validator against it, rather than trusting the docs. That's the kind of finding a dogfooding pass is supposed to produce.

Side effects: None — confirmed no other repo touched.

Frontend: N/A.

Recommend: ready to merge, after #31.

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud