Refactor Slack notifications: severity-aware, human-centered format

[devin-ai-integration]

Refactor Slack notifications: severity-aware, human-centered format

Summary

Replaces the flat title + bullet list + context Slack message format with a structured, severity-aware layout. Drift notifications now have 4 clearly labeled sections (Change Summary, Impact, Docs Status, Action) and a severity-colored header (🟢 Non-Breaking, 🟡 Potentially Breaking, 🔴 Breaking).

slack_notify.py was rewritten:

• send_slack_notification() removed; replaced by send_drift_notification(DriftMessage) and send_audit_notification(title, summary, context)
• New build_drift_message() constructs a DriftMessage dataclass from the drift report's endpoint/schema lists
• Severity is inferred from the drift data for formatting purposes only (no detection/classification logic changed)
• Shared _post_or_print() handles webhook posting or stdout fallback

guard.py callers updated:

• cmd_check now passes structured report data through build_drift_message() → send_drift_notification()
• cmd_audit now uses send_audit_notification() with a calmer title (no 🚨)

Review & Testing Checklist for Human

• Severity inference is string-match-based and fragile: infer_severity() checks for "lost field" and "gained field" substrings in schema_changes entries. These strings are generated by drift_detect.py:DriftReport.as_bullet_list(). If the wording there ever changes, severity silently falls through to POTENTIALLY_BREAKING instead of the correct level. Verify you're comfortable with this coupling, or whether severity should be a first-class property of DriftReport instead.
• send_slack_notification() was deleted: This is a breaking API change. Confirm nothing else imports it — check CI workflow, pre-commit hook scripts, and any test files that might reference the old function name.
• tier3_drafted is accepted but never passed as True: build_drift_message() has a tier3_drafted param but guard.py never sets it. The "Migration guidance draft added" line is unreachable. Decide whether to remove it or wire it up.
• Audit notification tone changed: Title went from "🚨 API Call Guard: Runtime mismatches detected" to "Runtime Audit | API call mismatches detected" and the :warning: Slack emoji was dropped from context. Verify this is the tone you want for audit alerts.

Test Plan

1. Start server: uvicorn app.main:app --host 0.0.0.0 --port 8000 --reload
2. Save baseline: python -m drift_guard.guard baseline
3. Simulate non-breaking drift (add optional field): modify baseline to remove role field, run check, verify Slack payload shows 🟢 Non-Breaking with "No action required"
4. Simulate breaking drift (remove endpoint): modify baseline to add fake endpoint, run check, verify Slack payload shows 🔴 Breaking with "Review required before merge"
5. Test audit: make bad API calls, run python -m drift_guard.guard audit, verify Slack payload uses the new audit format
6. Verify CI still passes (no import errors from removed send_slack_notification)

Notes

• Tested locally end-to-end — GREEN and RED drift messages match the spec exactly
• Ruff lint passes
• Link to Devin run: https://app.devin.ai/sessions/23d1e596e5f845ec8c47886363e9b2b8
• Requested by: @justinemach

---

[devin-ai-integration]

🤖 Devin AI Engineer

I'll be helping with this pull request! Here's what you should know:

✅ I will automatically:

• Address comments on this PR. Add '(aside)' to your comment to have me ignore it.
• Look at CI failures and help fix them

Note: I can only respond to comments from users who have write access to this repository.

⚙️ Control Options:

• Disable automatic comment and CI monitoring

[devin-ai-integration]

✅ Devin Review: No Issues Found

Devin Review analyzed this PR and found no potential bugs to report.

View in Devin Review to see 5 additional findings.