ledger: record approved compile builder phase and review handoff

campaigns.yaml

@@ -74,8 +74,8 @@ campaigns:
   updated: '2026-10-13'
   next: 'Timing-noise audit implementation complete: #410, #411, #412 merged 2026-10-13 in order (4abe4226,
     3cf6b2ce, 6108d029); issue #362 closed. Final inventory 35 SOUND including caveats / 3 FRAGILE / 0 UNSAFE-SILENT.
-    No fresh measurements or scientific acceptance. Next priority: resolve compile-builder capability and freeze
-    setup-baseline choices. C1/C11 sample counts, C5 estimators, GPU memo interpretation and fresh repeats
+    No fresh measurements or scientific acceptance. Compile-builder phase #413 is implemented in open PR #414; setup-baseline
+    choices remain unfrozen. C1/C11 sample counts, C5 estimators, GPU memo interpretation and fresh repeats
     remain explicit domain follow-ups.'
   evidence: https://github.com/PyAutoLabs/autolens_profiling/pull/412
 - id: autofit
@@ -138,8 +138,8 @@ tasks:
   updated: '2026-10-13'
   next: 'Timing-noise audit implementation complete: #410, #411, #412 merged 2026-10-13 in order (4abe4226,
     3cf6b2ce, 6108d029); issue #362 closed. Final inventory 35 SOUND including caveats / 3 FRAGILE / 0 UNSAFE-SILENT.
-    No fresh measurements or scientific acceptance. Next priority: resolve compile-builder capability and freeze
-    setup-baseline choices. C1/C11 sample counts, C5 estimators, GPU memo interpretation and fresh repeats
+    No fresh measurements or scientific acceptance. Compile-builder phase #413 is implemented in open PR #414; setup-baseline
+    choices remain unfrozen. C1/C11 sample counts, C5 estimators, GPU memo interpretation and fresh repeats
     remain explicit domain follow-ups.'
   issue: https://github.com/PyAutoLabs/autolens_profiling/issues/362
   evidence: https://github.com/PyAutoLabs/autolens_profiling/pull/412
@@ -185,13 +185,16 @@ tasks:
 - id: jax_compile_probe_needs_own_cell_builder
   campaign: measurement-tools
   title: jax_compile/probe.py lost its cell builder with the searches tier — give profiling its own
-  status: ready
-  priority: normal
+  status: active
+  priority: high
   path: tasks/jax_compile_probe_needs_own_cell_builder.md
   updated: '2026-10-13'
-  next: Unchanged and unissued; scripts/misc/jax_compile/probe.py still on main (run state unverified). Select
-    one bounded step when prioritised.
-  evidence: https://github.com/PyAutoLabs/autolens_profiling/blob/a93f37a7ade16fcb7673a7777dfde505183728ed/scripts/misc/jax_compile/probe.py
+  next: 'Bounded phase approved and issued as autolens_profiling#413: explicit current-runtime imaging/MGE/HST
+    fp64 jit workload with isolated output, no historical-pin equivalence. PR #414 is open for review; local
+    validation and exact-head GitHub checks passed (2026-10-13); independent review CLEAN at implementation head 05c5c49. Await human merge command.
+    No production compute authorized. After review, agree a separate measurement plan or next capability;
+    global compile dispatch and broader builder coverage remain unavailable.'
+  evidence: https://github.com/PyAutoLabs/autolens_profiling/pull/414
 - id: mass_field_flat_adoption_science_repos
   campaign: measurement-tools
   title: 'Remaining code consumers move to the flat `fields=` form: profiling, inference, JOSS benchmarks,

dashboard.md

@@ -1,6 +1,6 @@
 # PyAutoPulse — profiling dashboard
 
-<!-- pulse-campaigns:e64626fe7153e68337cbc03ac09f26b3b0090c1f7ae885c38ddc562c69bc5955 -->
+<!-- pulse-campaigns:83d03e8f75348c47adfd20f3dca66dab49e6482bbee5acea7f30a713bba9c587 -->
 ## Profiling Check In
 
 <details><summary>Full check-in prompt</summary>
@@ -133,7 +133,7 @@ Reviewed 2026-10-13. No new production-over-sampling or post-certified measureme
 
 <details><summary>Measurement reliability and profiling tools — active</summary>
 
-Reviewed 2026-10-13. Timing-noise audit implementation complete: #410, #411, #412 merged 2026-10-13 in order (4abe4226, 3cf6b2ce, 6108d029); issue #362 closed. Final inventory 35 SOUND including caveats / 3 FRAGILE / 0 UNSAFE-SILENT. No fresh measurements or scientific acceptance. Next priority: resolve compile-builder capability and freeze setup-baseline choices. C1/C11 sample counts, C5 estimators, GPU memo interpretation and fresh repeats remain explicit domain follow-ups.
+Reviewed 2026-10-13. Timing-noise audit implementation complete: #410, #411, #412 merged 2026-10-13 in order (4abe4226, 3cf6b2ce, 6108d029); issue #362 closed. Final inventory 35 SOUND including caveats / 3 FRAGILE / 0 UNSAFE-SILENT. No fresh measurements or scientific acceptance. Compile-builder phase #413 is implemented in open PR #414; setup-baseline choices remain unfrozen. C1/C11 sample counts, C5 estimators, GPU memo interpretation and fresh repeats remain explicit domain follow-ups.
 
 [Campaign evidence](https://github.com/PyAutoLabs/autolens_profiling/pull/412)
 
@@ -143,7 +143,7 @@ Reviewed 2026-10-13. Timing-noise audit implementation complete: #410, #411, #41
 - [A gradient-cost probe: forward vs `value_and_grad` ms/eval and a strict FD check, on any registry cell](https://github.com/PyAutoLabs/PyAutoPulse/blob/main/tasks/gradient_cost_probe.md) — ready: Unissued. New context 2026-10-07: Heart unit-test timing flagged PyAutoLens forward-gradient tests +150%/+96% after PyAutoLens#768 rewrote that test file; not a controlled comparison, and this probe is the right instrument. Select one bounded step when prioritised.
 - [Numba breakdown harness: perturb the instance so the operated-matrix memo cannot hide a step](https://github.com/PyAutoLabs/PyAutoPulse/blob/main/tasks/numba_breakdown_harness_memo_blind.md) — ready: Unchanged and unissued (PyAutoArray#496 closed 2026-08-27). Select one bounded step when prioritised.
 - [Search settings-estimation + profiling infrastructure (n_starts / batch_size / n_batch)](https://github.com/PyAutoLabs/PyAutoPulse/blob/main/tasks/search_settings_estimation_infrastructure.md) — ready: Unchanged and unissued (autolens_profiling#82 closed 2026-08-18). Select one bounded step when prioritised.
-- [jax_compile/probe.py lost its cell builder with the searches tier — give profiling its own](https://github.com/PyAutoLabs/PyAutoPulse/blob/main/tasks/jax_compile_probe_needs_own_cell_builder.md) — ready: Unchanged and unissued; scripts/misc/jax_compile/probe.py still on main (run state unverified). Select one bounded step when prioritised.
+- [jax_compile/probe.py lost its cell builder with the searches tier — give profiling its own](https://github.com/PyAutoLabs/PyAutoPulse/blob/main/tasks/jax_compile_probe_needs_own_cell_builder.md) — active: Bounded phase approved and issued as autolens_profiling#413: explicit current-runtime imaging/MGE/HST fp64 jit workload with isolated output, no historical-pin equivalence. PR #414 is open for review; local validation and exact-head GitHub checks passed (2026-10-13); independent review CLEAN at implementation head 05c5c49. Await human merge command. No production compute authorized. After review, agree a separate measurement plan or next capability; global compile dispatch and broader builder coverage remain unavailable.
 - [profile_lens_aggregator.py cannot run from the autolens_workspace_developer root: no config/ directory](https://github.com/PyAutoLabs/PyAutoPulse/blob/main/tasks/profile_lens_aggregator_needs_config_dir.md) — ready: Unchanged: workspace_developer main still has no root config/ (bug presumed live, not re-run). Select one bounded step when prioritised.
 - [jax_profiling/gradient/imaging/pixelization.py: 3.2% of its pin move is unattributed](https://github.com/PyAutoLabs/PyAutoPulse/blob/main/tasks/gradient_pixelization_pin_residual_drift.md) — ready: Unchanged and unissued (PyAutoArray#490 merged 2026-08-26). Select one bounded step when prioritised.
 - [Pair JAX/XLA env vars with measured compile and run times, per backend](https://github.com/PyAutoLabs/PyAutoPulse/blob/main/tasks/pair_jax_xla_env_vars_with_measured.md) — ready: Unchanged and unissued. Context: the jax-version policy is now released (2026.10.7.1 libraries require Nerves with the JAX deadlock exclusions). Select one bounded step when prioritised.

tasks/jax_compile_probe_needs_own_cell_builder.md

@@ -17,6 +17,52 @@ The original task below is retained verbatim. Current scheduling state lives in
 - Evidence: [project main reviewed at 4db4d988](https://github.com/PyAutoLabs/autolens_profiling/blob/4db4d988e520a82219152743748c3188d485a4ba/wiki/campaigns/measurement_tools.md), [Mind claims at 35281b1d](https://github.com/PyAutoLabs/PyAutoMind/blob/35281b1d/active.md), [full dated review](../checkins/2026-10-13.md). Review date does not refresh measurement dates.
 - Availability: project `hpc/sync jobs` returned an empty current queue; `hpc/sync status` timed out after 25 seconds during the output preview. Historical job outcomes, unpulled results and RAL cleanup contents are unverified; no collection or compute was performed.
 
+
+## Approved bounded phase — 2026-10-13
+
+- Human approved the HST/MGE current-runtime builder plan in this conversation;
+  [implementation issue #413](https://github.com/PyAutoLabs/autolens_profiling/issues/413)
+  is active in Mind as `compile-builder-hst-mge`. This supersedes the earlier
+  "unissued" next step, without completing the broader domain task.
+- The original witness below is incompatible with current evidence: `--cell`
+  is not a supported CLI flag, and the retired flat-vector posterior differs
+  from the current model-instance-pytree runtime likelihood. Its pinned timings
+  cannot validate this restoration. The approved phase instead checks small
+  deterministic CPU objective parity and preserves historical results and pins.
+- Phase scope: imaging/MGE/HST, fp64, `jit`, existing data only; explicit objective
+  identity and isolated output. No gradient/batch/general-cell support, baseline
+  acceptance, production measurements or RAL/GPU submissions are authorized.
+- Verified locally: tiny CPU compiled/eager objective parity and unsupported-path,
+  import, output-isolation and legacy-consumer checks pass. These are implementation
+  tests, not campaign measurements. Brain's compile conductor still emits an empty
+  dispatch plan because the global builder capability remains unavailable.
+- Implementation: [PR #414](https://github.com/PyAutoLabs/autolens_profiling/pull/414),
+  head `05c5c49a514db1c3dc2b92d1126cf39d7f4cd68a`, is OPEN, not merged. Shared
+  builder [source](https://github.com/PyAutoLabs/autolens_profiling/blob/18b1ee4/scripts/misc/cell_builders/imaging_mge.py)
+  and [instrument contract](https://github.com/PyAutoLabs/autolens_profiling/blob/18b1ee4/scripts/misc/jax_compile/README.md)
+  were verified 2026-10-13. No production timing was collected.
+- Validation: 19 targeted tests; full suite 1,378 passed / 6 skipped, with all 45
+  baseline-only failures passing after reverting an inventory-hash-changing route
+  description edit. The registry and baseline remain byte-for-byte unchanged.
+  Project import smoke, lint, generated catalogue/dashboard and contract checks pass.
+  Exact-head [project CI](https://github.com/PyAutoLabs/autolens_profiling/actions/runs/38426218793)
+  passed 2026-10-13: 1,424 tests passed / 6 skipped, plus browser, link, smoke and
+  generated-contract checks. Pulse head `685e1b8` passed both
+  [lint](https://github.com/PyAutoLabs/PyAutoPulse/actions/runs/38426360330) and
+  [refresh dry run](https://github.com/PyAutoLabs/PyAutoPulse/actions/runs/38426360116).
+  Independent review was CLEAN at project `05c5c49` and Pulse `a76a769` on
+  2026-10-13: 40 focused tests passed; the real tiny-FITS CPU CLI path with a
+  3×3 PSF passed within the 60-second verification cap; baseline, registry and
+  historical files were unchanged. Temporary verification artifacts were removed.
+  These checks establish implementation validation, not timing improvement or
+  scientific acceptance. Both PRs still require the judge-tier human merge command.
+- Plan adjustment: partial capability is documented in the instrument and campaign
+  journal; the machine-readable global route reason is retained because the baseline
+  draft pins the registry's exact bytes. No re-freeze was authorized or performed.
+- Next: review and merge the implementation PR when authorized and green, then discuss
+  the next bounded experiment and resource request. Broader coverage and compatible
+  baselines remain pending. Review date does not refresh measurement freshness.
+
 ---
 
 # jax_compile/probe.py lost its cell builder with the searches tier — give profiling its own