Both were sitting untracked in the repo root. They're the source for the "candidate N from the 2026-06-10 review" references in this batch of follow-up PRs, so keeping them in-repo makes those references resolvable. Self-contained HTML; a short README indexes them. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
17 lines
812 B
Markdown
17 lines
812 B
Markdown
# Reviews
|
|
|
|
Point-in-time review decks. Self-contained HTML — open directly in a
|
|
browser (arrow keys / click to navigate). Each is a snapshot of its
|
|
date; findings were triaged and tracked through PRs, so the deck reads
|
|
as the "why" behind a batch of follow-up work rather than a live
|
|
checklist.
|
|
|
|
- **`architecture-review-2026-06-10.html`** — deepening opportunities
|
|
(shallow modules, implicit interfaces, duplicated choreography). The
|
|
10 candidates were implemented across PRs that reference them as
|
|
"candidate N from the 2026-06-10 review".
|
|
|
|
- **`security-review-2026-06-10.html`** — defensive review of the auth,
|
|
federation, secrets, and injection surfaces. Severity-calibrated
|
|
(several scanner false-positives were verified and dropped). All
|
|
actionable findings were fixed in follow-up PRs.
|