felt: sync fiber store

Backfill ULID ids across 19 fibers; close docs-versioning,
smoke-test-read-only, docker-uv-revert (superseded by #733);
refresh shapepipe.md active-threads list to current PRs (#737–741);
add np-str0-numpy2 fiber; minor outcome/status normalizations.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

Author: cailmdaley

.felt/dependabot-pr-triage/dependabot-pr-triage.md

@@ -1,4 +1,5 @@
 ---
+id: 01KTCHWZSHQXRQHTHH9Q7Q4MT7
 name: Triage open dependabot uv.lock PRs
 status: closed
 tags:

.felt/docker-multistage/docker-multistage.md

@@ -1,4 +1,5 @@
 ---
+id: 01KTCHWZV7263MYPNB8DT1FXPX
 name: 'Docker multi-stage: runtime + dev targets'
 tags:
     - shapepipe

.felt/docker-uv-revert/docker-uv-revert.md

@@ -1,12 +1,14 @@
 ---
+id: 01KTCHWZW8DT8VRYQJ4CEPDRGB
 name: 'Docker: revert skaha→python base, switch to uv lockfile'
-status: active
+status: closed
 tags:
     - shapepipe
     - docker
     - infra
 created-at: 2026-04-27T11:26:45.677512058+02:00
-outcome: 'PR #719 (chore: switch Dockerfile to slim Python + uv lockfile) opened and CI-green on first try (3m31s); ready for Martin''s review. Drops conda double-install, makes pyproject SSOT + uv.lock the pinned manifest, switches WeightWatcher from sed-patched source build to Debian''s pre-patched 1.12+dfsg-3 package, adds binary smoke tests to deploy-image.yml.'
+closed-at: 2026-06-10T17:14:45.965931602+02:00
+outcome: 'Superseded: conda removal landed via #733 (merged); the #719-era docker-uv work is absorbed into ci-green-on-develop, which tracks remaining follow-ups.'
 decisions:
     base:
         label: Base image

.felt/fabian-coord-bug/fabian-coord-bug.md

@@ -1,4 +1,5 @@
 ---
+id: 01KTCHWZWEXTJ337APFSH4NF6X
 name: Fabian's coord-propagation bug + image-sim code on github
 tags:
     - shapepipe

.felt/ngmix-update/ngmix-update.md

@@ -1,4 +1,5 @@
 ---
+id: 01KTCHWZWY48FVCEDX6DGE6CTY
 name: ngmix library upgrade + Lucy wrapper sync
 tags:
     - shapepipe

.felt/prs-in-flight/prs-in-flight.md

@@ -1,4 +1,5 @@
 ---
+id: 01KTCHWZWYYTDSYHAP4FPE7V3V
 name: PRs in flight after v2 merge
 tags:
     - shapepipe

.felt/review-ngmix-v2-pr740/review-ngmix-v2-pr740.md

@@ -1,22 +1,23 @@
 ---
 id: 01KTCHWZX873VG28AA663A3ZQE
 name: 'Review + work: ngmix v2.0 (PR #740)'
-status: active
+status: closed
 tags:
     - constitution
     - shapepipe
     - ngmix
     - review
 created-at: 2026-06-01T11:50:17.967872934+02:00
+closed-at: 2026-06-05T23:59:38.502Z
 outcome: 'INTERACTIVE — live with Cail. Bucket-A cleanups PUSHED to ngmix_v2.0/#741 (dd4f656a..bd60dc8e: 293/1140/254-resume/766 + fitting.py deleted); CI running. Dev image built (/n17data/cdaley/containers/shapepipe-dev) + ngmix 2.4.0 installed → full suite + example pipeline now runnable. Ran a 6+2-agent WORKFLOW on the two hard problems → two decision-ready reports in fiber: weights-report.md, size-report.md, deep-dive-report.html (sent to Cail). WEIGHTS (#604+949): found TWO coupled regressions in prepare_ngmix_weights (ngmix.py:871) — R1 noise estimator regressed from object-free windowed get_noise (now DEAD at :826) to flux-contaminated whole-stamp sigma_mad; R2 lost the v1 binarization → double-counts a real inverse-variance map. EMPIRICALLY CONFIRMED (truth ivar 1e6 → recovered 8.8e11, ratio≈1/noise²). Clean red→green test target via make_data. Rec: SPLIT — minimal v1-restore (reinstate binarization + get_noise) + test in #741; SExtractor BACKGROUND_RMS baseline as a SEPARATE PR (Codex shuttle, ~4-8h, closes #604). SIZE (r50/T): galaxy r50=pars[4]=T (area), PSF r50psf=σ; neither is 1.1774σ; all 5 r50* cols new in v2.0. UNIONS-3500 WL I (arXiv:2605.13549) reports half-light radius r_h as PRIMARY. Rec: TRANSFORM at source (honest r50 in ngmix) + cs_util converter web; bonus find — sp_validation galaxy.py:T_to_fwhm is dimensionally wrong (feeds the scale-dependent PSF-leakage fit). Both work-streams now DISPATCHED as shuttles (Cail''s go): [[ngmix-weights-ivar]] (Codex — minimal #741 regression fix w/ red→green test + the SExtractor BACKGROUND_RMS baseline as a separate PR closing #604) and [[ngmix-size-columns]] (claude-opus — honest r50 at the ngmix source + cs_util converter web + the sp_validation T_to_fwhm leakage fix). Workers prepare branches/commits/PR-descriptions + reports; merge/PR-creation stays Cail''s. STILL OPEN for the eventual #741 reply (not in these streams): *_psfo reconvolved-PSF merge-gate (1045) + the runner-decorator contract (testable via the example pipeline). PLAN: let the two shuttles finish → reassess → write the Martin reply. No reply drafted yet.'
 horizon: now
 shuttle:
-    agent: claude-opus
     enabled: true
-    host: candide
-    interactive: true
     kind: oneshot
+    interactive: true
+    host: candide
     project_dir: /automnt/n17data/cdaley/unions/shapepipe
+    agent: claude-opus
 ---
 
 # Review + work: ngmix v2.0 (PR #740)

.felt/shapepipe.md

@@ -1,4 +1,5 @@
 ---
+id: 01KTCHX00NMQ1VDPGXRYJ6RGZR
 name: ShapePipe maintenance & PRs
 tags:
     - shapepipe
@@ -26,17 +27,20 @@ Surfaced over a 2026-04-27 walking conversation. Captured in
   Disclosure on Claude-only review per
   `feedback_claude_only_review_disclosure`.
 
-## Active threads
+## Active threads (refreshed 2026-06-10)
 
-- **[[shapepipe/docker-uv-revert]]** — slim Python + uv lockfile, drop conda. PR #719 (draft).
-- **[[shapepipe/prs-in-flight]]** — tracking #708 (testing scaffold), #714 (develop bugs), #719 (this one).
+- **PR #737** (MPI on candide + containerized SLURM scripts) — **MERGED 2026-06-10** (rebased onto develop, both CI runs green).
+- **PR #738** (versioned docs + switcher) — **MERGED 2026-06-10** after independent review; sfarrens offered post-hoc comments. Stable root refreshes on next master push. See [[docs-versioning]].
+- **PR #739** (machine-specific cluster docs tree) — awaiting Martin's review; check rebase state after the #737/#738 merges. See [[docs-cluster-tree]].
+- **PR #741** (ngmix v2.0, CI mirror of #740) — Martin left 10 inline comments Jun 5, no verdict yet; shear recovery verified unbiased to ~1e-4 in m. See [[ngmix-weights-ivar]] (PR-ready regression fix + #604 ivar plan) and [[ngmix-size-columns]] (honest r50 spec).
+- **Martin's PRs #704 (contributors) & #699 (coverage mask)** — Cail's review requested; both conflicting, need Martin's rebase first.
+- **[[ci-green-on-develop]]** — conda fully removed (#733 merged); remaining follow-ups tracked there.
 
-## Future work
+## Earlier threads (superseded)
 
-- **[[shapepipe/ngmix-update]]** — replace Axel's stable_version fork
-  with upstream ngmix; reconcile with Lucy's wrapper.
-- **[[shapepipe/fabian-coord-bug]]** — port Fabian's 1-line coord
-  propagation fix; first need his image-sim code on github.
+- [[shapepipe/docker-uv-revert]] / [[shapepipe/prs-in-flight]] — the #719-era conda removal; landed via #733, current state in [[ci-green-on-develop]].
+- **[[shapepipe/ngmix-update]]** — became the #740/#741 ngmix v2.0 review thread.
+- **[[shapepipe/fabian-coord-bug]]** — port Fabian's 1-line coord propagation fix; still pending his image-sim code reaching github.
 
 ## Conventions specific to this repo
 

.felt/shapepipe/ci-develop-trigger/ci-develop-trigger.md

@@ -1,4 +1,5 @@
 ---
+id: 01KTCHWZX8GRCMJHGCRARBGFS8
 name: CI silently broken on develop; install_shapepipe conda.sh lookup
 status: closed
 tags:

.felt/shapepipe/ci-green-on-develop/ci-green-on-develop.md

@@ -1,4 +1,5 @@
 ---
+id: 01KTCHWZX9Q1MG2FB20N5Y52TD
 name: 'ShapePipe CI: green & trustworthy test suite on develop'
 status: open
 tags: