diff --git a/active.md b/active.md
index a23155e6..20e9817e 100644
--- a/active.md
+++ b/active.md
@@ -1,50 +1,5 @@
# Active Tasks
-## latex-raw-string-docstrings
-- issue: https://github.com/PyAutoLabs/autolens_workspace/issues/491
-- status: workspace-dev — 5 of 6 PRs MERGED; HowToLens#73 held RED (not this PR's failure)
-- issue-umbrella: autolens_workspace#491 (one issue, six PRs)
-- merged: HowToFit#47, HowToGalaxy#70, autofit_workspace#145, autogalaxy_workspace#218,
- autolens_workspace#492 — all green, incl. the CI 'Catalogue staleness' check, which
- independently re-confirms the regeneration gate.
-- BLOCKED: HowToLens#73 red on smoke (3.12 + 3.13). Two scripts fail with numba
- 'Pass nopython_type_inference': chapter_3_pixelizations/tutorial_8_adaptive_pixelization.py
- and tutorial_11_brightness_adaption.py. NEITHER is in the diff, and all 8 files that ARE
- in it pass. Reproduced identically on one re-run. A docstring r-prefix cannot cause a numba
- typing error. Prime suspect: PyAutoArray#453 (in-place Cholesky buffer + new numba kernels
- for fnnls_cholesky) merged 2026-08-20 22:09:30 UTC, minutes before this run; CI installs
- autoarray from source at HEAD, and the last green HowToLens main run (04:29 UTC) predates
- it. NOT yet proven: two specific mechanisms were tested and BOTH passed under the exact CI
- versions (numba 0.67.0 + scipy 1.17.1) — a strided Ubuf view into _cholupdate, and the new
- np.dot inside _cho_solve_buffer. So the suspicion rests on timing + code area, not a
- reproduced mechanism. Control experiment running: re-run of the last GREEN main build
- (same commit, fresh dependency install) — if it now fails the same way, the break is on
- main/library drift and this PR is clean.
-- prompt: active/latex_raw_string_docstrings.md
-- repos (all on branch claude/latex-raw-string-docstrings-9h4ine):
- - HowToFit: 4 files, 13 literals, 7 corruptions repaired
- - HowToGalaxy: 4 files, 20 literals, 13 repaired
- - HowToLens: 8 files, 32 literals, 21 repaired
- - autofit_workspace: 2 files, 2 literals, 1 repaired
- - autogalaxy_workspace: 6 files, 30 literals, 28 repaired
- - autolens_workspace: 17 files, 83 literals, 61 repaired
-- totals: 41 files, 180 literals, 131 corruptions repaired — 41 matches the survey exactly.
-- verification (per repo, in order): baseline regeneration is a NO-OP; both sweeps zero;
- runtime value check (every changed literal's value compared HEAD vs worktree — the prefix
- may only REMOVE corruption) = 131 repaired / 0 unexpected; diff-empty gate passes
- byte-exactly on notebooks/ markdown/ llms-full.txt workspace_index.json. autolens_workspace
- additionally: all 57 __Env__ declarations re-read IDENTICALLY, and 4 of the raw-stringed
- files carry __Env__, so the PyAutoHands#251 fix is exercised end to end.
-- gate refinement: HowToFit's 4 `plt.ylabel` labels are runtime strings in CODE cells, which
- copy source verbatim, so the `r` legitimately shows there. The gate holds exactly as written
- for every docstring; that one code-cell delta is by design.
-- residue: autolens_workspace scripts/group/likelihood_function.py keeps 2 warned + 1 silent.
- Three docstrings use the DOUBLE-backslash convention ($\\theta$, \\frac, \\vec) mixed with
- single-backslash macros; adding `r` would double the correct ones. Needs a convention
- decision + un-doubling 18 backslashes = a prose edit this task excludes. Follow-up.
-- follow-up to file: -W error::SyntaxWarning compile guard in workspace CI (catches only the
- warned class; the silent class needs the AST sweep).
-
## numba-cpu-likelihood-profiling
- issue: https://github.com/PyAutoLabs/autolens_profiling/issues/151
- pr: https://github.com/PyAutoLabs/autolens_profiling/pull/152
diff --git a/active/latex_raw_string_docstrings.md b/complete/2026/08/latex-raw-string-docstrings.md
similarity index 67%
rename from active/latex_raw_string_docstrings.md
rename to complete/2026/08/latex-raw-string-docstrings.md
index cae959a4..36308b63 100644
--- a/active/latex_raw_string_docstrings.md
+++ b/complete/2026/08/latex-raw-string-docstrings.md
@@ -1,3 +1,75 @@
+- issue: https://github.com/PyAutoLabs/autolens_workspace/issues/491 (one issue, six PRs)
+- shipped: 2026-08-20 — all six merged:
+ HowToFit#47, HowToGalaxy#70, HowToLens#73, autofit_workspace#145,
+ autogalaxy_workspace#218, autolens_workspace#492.
+- classification: maintenance (6 workspace repos) — prose-only, no API surface.
+- blocked-by (cleared): hands-raw-string-docstring-prefix / PyAutoHands#251.
+- summary: prefixed `r` on every LaTeX-carrying docstring across the six workspace repos —
+ **41 files, 180 literals, 131 silent corruptions repaired**. 41 files matches the
+ 2026-08-20 survey exactly, and autolens_workspace's 17 files / 80 warnings match the
+ independent 2026-08-09 measurement line for line. Per repo: HowToFit 4/13/7,
+ HowToGalaxy 4/20/13, HowToLens 8/32/21, autofit_workspace 2/2/1,
+ autogalaxy_workspace 6/30/28, autolens_workspace 17/83/61 (files/literals/repaired).
+- the two damage classes: `SyntaxWarning` fires only for escapes Python does NOT recognise
+ (`\s`, `\l`, `\[`). The ones it DOES recognise — `\t` in `\theta`, `\f` in `\frac`,
+ `\r` in `\rm`, `\b` in `\beta` — corrupt the value with NO diagnostic at all.
+ `\theta_E` in HowToLens chapter_4 was literally TAB + "heta_E". Confirmed repaired.
+- key traps:
+ - **THE DIFF-EMPTY GATE CANNOT CATCH A WRONG EDIT HERE.** The prompt's gate (regenerate,
+ assert notebooks/markdown/llms-full.txt/workspace_index.json unchanged) is necessary but
+ structurally blind: the generator reads SOURCE TEXT, not runtime values, so a corrupted
+ value regenerates byte-identically. The check that actually catches mistakes is comparing
+ every changed literal's RUNTIME VALUE HEAD vs worktree and asserting the `r` prefix only
+ ever REMOVES corruption. It caught two real errors mid-task: a `print("\nInfo:")` newline
+ about to be destroyed, and 12 already-escaped `$\\chi^2$` about to be doubled into LaTeX
+ line breaks. Final tally: 131 repaired, 0 unexpected. **Build this check first.**
+ - **`\\` is ALWAYS deliberate, whatever the surrounding context.** In a non-raw literal
+ `\\` already means one literal backslash — the author got it right; adding `r` doubles it.
+ A math-context rule alone lets these through (they sit inside `$...$`).
+ - **The warning sweep is interpreter-dependent.** Invalid escapes are a `SyntaxWarning`
+ only on Python 3.12+; on 3.11 they are a `DeprecationWarning`. Verified on 3.11.15: a
+ docstring with `$\odot$` yields 0 SyntaxWarning and 1 DeprecationWarning — so a
+ SyntaxWarning-only sweep returns a VACUOUS zero that looks exactly like "already fixed".
+ Collect both. (`compileall` also needs `-f`, or `__pycache__` hides the counts.)
+ - **Run the generator BEFORE editing and confirm a no-op.** Otherwise pre-existing
+ generator drift is indistinguishable from damage your edit caused.
+ - **Preserve line endings.** Several scripts are CRLF; a `write_text` round-trip normalises
+ the whole file and turns a 1-character edit into a 549-line diff. Use `newline=""`.
+ - Auto-fix only where EVERY backslash sits in a LaTeX context — `$...$`, `\(...\)`,
+ `\[...\]`, `\begin{}...\end{}`, or a markdown code span. Four docstrings with malformed
+ or unbalanced delimiters were read in full and prefixed by hand; their delimiters were
+ left exactly as found (fixing them is a prose change this task excluded).
+ - **Gate refinement**: runtime label strings (e.g. HowToFit's four `plt.ylabel`) live in
+ CODE cells, which copy source verbatim, so the `r` legitimately appears in the notebook.
+ The gate holds exactly as written for every docstring; that one delta is by design.
+- residue (deliberate, worth a follow-up): `autolens_workspace/scripts/group/likelihood_function.py`
+ keeps 2 warned + 1 silent. Three of its docstrings use the DOUBLE-backslash convention
+ (`$\\theta$`, `\\frac`, `\\vec`) mixed with single-backslash macros; adding `r` would double
+ the already-correct ones, and fixing it properly means un-doubling 18 backslashes — a prose
+ edit this task excluded. Needs a convention decision.
+- also left alone: already-escaped `\\` anywhere (two autogalaxy interferometer files,
+ autofit `samples.py:410`), real newlines in `print()`, and
+ `autolens_workspace/dataset/cluster/a2744/prep.py:38` where `line.split("\t")` is a genuine
+ TSV tab.
+- **HowToLens#73 was merged with smoke RED, and the red was proven not to be this PR's.**
+ Two chapter_3 pixelization scripts fail numba `Pass nopython_type_inference`; neither is in
+ the diff and all 8 files that are in it pass. Decisive control: re-running the last GREEN
+ `main` build (run 255, commit `dcb67e9`, zero changes from the branch, fresh dependency
+ install) now FAILS identically — `main` itself is red. Suspected cause, NOT proven:
+ PyAutoArray#453 (in-place Cholesky buffer + new numba kernels for `fnnls_cholesky`) merged
+ 22:09:30 UTC, minutes before. Two candidate mechanisms were tested under the exact CI
+ versions (numba 0.67.0 + scipy 1.17.1) and BOTH passed — a strided `Ubuf` view into
+ `_cholupdate`, and the new `np.dot` in `_cho_solve_buffer`. A new numba release was ruled
+ out (0.67.0 shipped nine days before the last green run). **Open question for whoever owns
+ that solver.**
+- follow-up shipped: PyAutoBrain#245 added a `hygiene escapes` mode so this debt is caught
+ continuously — it reports both classes, marks files with ONLY silent damage, and handles
+ both interpreter traps. Preferred over the originally-proposed
+ `-W error::SyntaxWarning` CI guard, which would catch only the warned class and pass
+ vacuously on 3.11.
+
+## Original prompt
+
# Raw-string the LaTeX docstrings across the six workspace repos
Type: maintenance
diff --git a/complete/index.md b/complete/index.md
index bc56e428..b0f18cdb 100644
--- a/complete/index.md
+++ b/complete/index.md
@@ -6,7 +6,7 @@ Token-light navigation over the finished-work records (schema:
only then grep a dated bucket. Curators: edit the band between the CURATED
markers; everything below GENERATED is rebuilt.
-1061 records across 7 buckets.
+1062 records across 7 buckets.
## Highlights
@@ -93,6 +93,7 @@ _(curate hard-won records here — survives regeneration.)_
- [jax-likelihood-mass-sensitivity](2026/08/jax-likelihood-mass-sensitivity.md) — Audit + fix for median-literal mass blindness across all jax_likelihood/datacube scripts (phase 2 of the findi…
- [knowledge-board](2026/08/knowledge-board.md) — auto-closed on merge
- [lacosmic-cr-option-and-star-pass-decoupling](2026/08/lacosmic-cr-option-and-star-pass-decoupling.md) — both deliberately OPEN — closure is the default-flip decision, not this ship
+- [latex-raw-string-docstrings](2026/08/latex-raw-string-docstrings.md) — one issue, six PRs
- [lazy-heavy-imports](2026/08/lazy-heavy-imports.md) — Deferred all heavy non-essential imports to first use. `import autolens` 4.07s → 1.2–1.3s
- [llms-txt-census-fixes](2026/08/llms-txt-census-fixes.md)
- [markdown-renderings-2a-leftovers](2026/08/markdown-renderings-2a-leftovers.md)
diff --git a/dashboard.html b/dashboard.html
index 7de4b0ee..f65f7a31 100644
--- a/dashboard.html
+++ b/dashboard.html
@@ -38,7 +38,7 @@
📋 PyAutoMind Dashboard
Every task the Mind is holding. Tap a task's 📋 and its /start_dev command is on your clipboard — paste it into a Claude Code chat to route Claude straight to that task.
-In flight 6 · Parked 3 · Planned 6 · Backlog 145 · markdown version
+In flight 5 · Parked 3 · Planned 6 · Backlog 145 · markdown version
Start here
Highest priority (filed as high) — showing 12 of 33
@@ -60,7 +60,6 @@