I swept /Users/samaguiar/Documents/SAIL and its child trees. I verified there are no broken symlinks, nested Git repos, empty folders, duplicate/copy-suffixed files, obvious backup artifacts, or credential-shaped files inside the SAIL tree. The JSON state files all parse cleanly.
I removed two safe local hygiene files: /Users/samaguiar/Documents/SAIL/.DS_Store and /Users/samaguiar/Documents/SAIL/slop-state/.DS_Store, then verified no .DS_Store files remain under SAIL.
I created the local report at /Users/samaguiar/Documents/SAIL/qa-sweeps/sail-folder-sweep-2026-05-20.md.
The SAIL-local May 17 structural state still shows 516 failing URLs and 17 passing URLs after the prior auto-fix pass. Remaining categories include 477 word-count-under-floor, 284 sidebar phone, 195 banned FAQ, 95 no-modern-components, 64 TOC-title issues, 21 hero 404s, 1 non-collapsible TOC, and 1 banned-heading issue.
The May 17 slop state still has 258 URLs in rewrite-queue.json. The rewrite folder only contains the README; the 24 individual JSON payloads referenced as /tmp/slop-work/rewrites/ are no longer present. The README marks 7 of the first 25 candidates as PUSH, but those still need humanizer plus brand QA because the first-pass transform was deterministic regex and can create grammar issues.
A newer audit from May 20 exists at /Users/samaguiar/Documents/Projects/wp-content-visual-scripts-2026/outputs/recent-wp-content-audit-2026-05-20/. It audited 90 recent WordPress items and found 0 pass, 20 warn, 70 fail. It is the better current repair queue for recent WordPress drift, while the SAIL May 17 state remains useful for full-site backlog context.
I also found the active Codex automation daily-website-content-drift-qa is configured to cwd /Users/samaguiar/Documents/New project, but that folder does not exist. The real script and latest outputs are under /Users/samaguiar/Documents/Projects/wp-content-visual-scripts-2026/. This is a remaining routine fix.
Outside SAIL, the hard-linked mirror files point into /Users/samaguiar/Documents/.tmp.driveupload/, which contains 50,377 files and uses about 60G. I did not clean it because it is outside the requested tree and should get a rollback-safe disk cleanup pass.
A future agent can start with the local report, then repair the automation cwd, prioritize the May 20 published-page failures, handle the 21 hero 404s from the full-site SAIL state, regenerate or rerun the missing slop rewrite payload export, and separately inspect .tmp.driveupload before any deletion.
No live WordPress changes were made in this session. The only file removals were .DS_Store files inside SAIL. The new report is the main local handoff artifact. If continuing the repair lane, use the May 20 audit for current recently edited content, not only the May 17 SAIL snapshot. The SAIL snapshot is still valuable for full-site backlog and historical state.
The requested sweep and QA/fix inventory are complete. Remaining work involves scheduled automation repair, live WordPress repair batches, missing rewrite payload regeneration, and outside-scope temp-upload cleanup.
Importance: High. Status: pending follow-up, because the folder sweep is complete but the QA backlog and automation-path fix remain open.