[campaign #252] STEP 5 — TAIL BATCH: smallest tail-end repos (≤20 files × ~50 repos)

[hyperpolymath]

Part of #252. Blocked by step 4.

Goal

Port the long-tail of small repos first. Per umbrella scale summary:

• ~50 repos in the ≤20-files-each tier
• Per-repo PR strategy (one PR per repo, per the TS-campaign convention)
• Each port: surface any residual stdlib gap, fold back into STEP 4 if any new bindings needed

Strategy notes (from umbrella body)

Realistic estimate: 200-400 PRs over many months.
Cannot use "smallest-first" naively — idaptik and panll dominate.
Better strategy: per-repo or per-subsystem batches.

Tail-end means per-repo batches at this tier; mid-tier (STEP 6) and large (STEP 7) may use per-subsystem.

Per-PR ownership gate

gh repo view hyperpolymath/<repo> --json owner,isFork,parent inline before each PR opens.

Acceptance

• ~50 per-repo PRs landed (each in scope of one tail repo)
• Ownership gate passed on every PR
• Residual stdlib gaps either folded back to STEP 4 or documented
• Memory tracker updated with per-repo completion ticks

Sequencing

Blocks: nothing in mainline sequence directly (STEP 6 can proceed in parallel with the tail of STEP 5 if mid-tier stdlib coverage is stable). Realistically: bulk of estate file-count surface still upstream in STEPS 6-8.

[hyperpolymath]

Re-measured 2026-08-27 — the tail is 12 repos / 49 files, not ~50 repos

	claimed	measured
repos with 1–20 .res	~50	12
files in the tier	—	49
repos with any .res estate-wide	—	15 of 420

Tail: metadatastician/vordr 15 · rescript 8 · social-media-tools 7 · jtv-lang 5 · boj-server 5, then singles and doubles.

⚠️ Caveat that changes how to read the shrinkage

The tier did not shrink purely by completion. Part of it shrank by the deletion mechanism documented on #279 and #664: 43% of estate .affine files are sub-400-byte Harvard Engine stubs standing 1:1 against deleted sources.

Do not score "12 left" as ~76% done without checking history per repo. A .res count of zero is satisfied identically by a port and by a deletion.