humanish 0.76.0 → 0.78.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
@@ -3,7 +3,7 @@
3
3
  Date: 2026-06-02 (current-state note updated 2026-07-14)
4
4
 
5
5
  Status: reference map for the major contracts shipped through source version
6
- `0.76.0`; it is not an exhaustive inventory of command/result envelopes. Exported types,
6
+ `0.78.0`; it is not an exhaustive inventory of command/result envelopes. Exported types,
7
7
  schema constants, parsers, and validators in `src/` are authoritative. Rows
8
8
  marked "reserved" name layering intent only — no code emits or validates them
9
9
  yet. Do not emit a reserved schema.
@@ -29,7 +29,7 @@ study completed, reproduced, and produced a real accessibility finding via a
29
29
  keyboard-first participant
30
30
  ([docs/goals/email-gated-signup/receipts/](email-gated-signup/receipts/)).
31
31
 
32
- ## Current Program Truth (source `0.76.0`)
32
+ ## Current Program Truth (source `0.78.0`)
33
33
 
34
34
  The package source and repository implementation in this tree agree on these
35
35
  points:
@@ -463,11 +463,43 @@ Stop and correct course if:
463
463
 
464
464
  ## Best Next Work
465
465
 
466
+ **2026-09-04, later (0.78.0).** `@e2b/desktop` moved from 2.2.3 to 2.3.3 (#638): its 2.3.1
467
+ changelog names the socket #581 found today, a background command's event stream the SDK kept
468
+ open after `Xvfb` and `startxfce4` were launched, which held the CLI process alive twelve minutes
469
+ past a run's written result. `doctor` now prints the installed desktop SDK version on its row and
470
+ attaches an advisory with the fix below 2.3.1 (#639). A live try-live on the new SDK reached its
471
+ goal in 104 s and settled with no active resources.
472
+
473
+ **2026-09-04 (0.77.0).** Four fixes found by runs and one by a scanner. A phone-emulated lane now
474
+ follows the participant into a tab that opens later: the hold-mode applier attaches to every page
475
+ target Chrome creates, paused before its first navigation, so the new tab lays out at the phone
476
+ width, and the state observer reads each new tab's own fidelity report, recording a covered tab
477
+ under `desktopGeometry.fidelity.laterTargets` (width, DPR, touch points) and a drifted one as a
478
+ lane warning with the number the page gave; the applier's own log travels as `holderLog`. The
479
+ first live proof failed in the participant's hands (a paused popup never resumed) and the shipped
480
+ design is the one four live runs settled on: never pause a later tab, send the overrides the
481
+ moment it exists, reload it once after its first navigation commits
482
+ (`docs/goals/computer-use-actor/receipts/later-tab-emulation-2026-09-04.md`, #623, #636). A sandbox create
483
+ or archive upload that fails on a transient provider error is retried once and named on the lane
484
+ (#630): six lanes started 20 s apart lost five to E2B SDK errors before any turn, a probe of the
485
+ same SDK a minute later worked in 6 s. The provider's `Retry-After` now governs a retry (capped
486
+ at 60 s) and a `403 misalignment_policy_violation` ends the lane named and unresumed (#633). The
487
+ provisioned-clone CI flake runs on an injected clock (#276 closed). The fourth benchmark run on
488
+ this build read 15 of 15 planted defects with nothing invented in three clean runs, cumulative 58
489
+ of 60 (`bench/RESULTS-2026-09-04-0.76.0.md`). humanish.dev negotiates `Accept: text/markdown`
490
+ and llms.txt says when to use humanish (#634).
491
+
492
+ ## Best Next Work
493
+
466
494
  **2026-09-03, later (0.76.0).** A hosted Chromium lane on a mobile preset can now be a
467
495
  mobile-emulated browser: `execution.desktop.fidelity.mobileEmulation: true` applies a real 414 px
468
496
  CSS viewport, the preset's device pixel ratio, touch events and a mobile user agent before the
469
497
  participant's first observation, and the bundle records what the page then reported about itself
470
498
  under `desktopGeometry.fidelity` (#221, `docs/goals/computer-use-actor/receipts/mobile-emulation-2026-09-03.md`).
499
+ On the published 0.76.0, with touch emulated, neither phone participant on TodoMVC could rename at
500
+ all (a double-tap selects text; the double-click editor never opens), where the 500 px runs without
501
+ touch had finished and only called the rename touch-hostile: the responsive-viewport study reported
502
+ an opinion, the mobile-emulated study reported a blocker.
471
503
  The persona axis was replicated the same evening on drawDB and TodoMVC and given a third app,
472
504
  Excalidraw, as the clean control (`persona-axis-phone-2026-09-03.md`); a multi-lane study's
473
505
  second and third findings reach `feedback draft` through `--candidate` (#609); a negated report
@@ -489,7 +521,7 @@ gpt-6-astra is priced. #548 closed on the benchmark receipts.
489
521
 
490
522
  **2026-09-01 (0.66.0 through 0.70.0, one session).** The product got its first efficacy numbers and
491
523
  they are in the repo with run ids: planted-defect recall 43 of 45 over three benchmark runs and
492
- 12 clean runs with nothing invented (`bench/`), 5 of 6 findings confirmed on TodoMVC and 11 of 12
524
+ 12 clean runs with nothing invented (`bench/`; 58 of 60 and 15 clean runs after the 09-04 rerun), 5 of 6 findings confirmed on TodoMVC and 11 of 12
493
525
  on drawDB against the source (20 participants on apps we did not write, 0 false reports), and the
494
526
  persona question answered on two apps: every keyboard-first participant hit a defect no
495
527
  mouse-driving participant met (drawDB's mouse-only database modal, TodoMVC's double-click-only
@@ -504,9 +536,10 @@ and a Claude participant with no memory across turns (#520). `humanish stats` (#
504
536
 
505
537
  The standing queue, in rough order:
506
538
 
507
- 1. #581: after a failed terminal run the CLI process lingered sixteen minutes past its result;
508
- the lane is clean in-process, so the holder is the SDK or the shell (`HUMANISH_DEBUG_HANDLES=1`
509
- now names it on the next occurrence);
539
+ 1. #581, the remaining half: a `Sandbox.create` that fails after the API allocated the sandbox
540
+ (the desktop SDK's own startup) leaves a sandbox this run never learned the id of, reclaimed only
541
+ by the provider's timeout; the linger itself is the SDK's held command stream, closed by the
542
+ 2.3.3 bump (0.78.0), and `doctor` names an older SDK;
510
543
  2. #513: there is still nowhere to read the docs on humanish.dev (site copy is locked; a docs
511
544
  route needs the operator's sign-off);
512
545
  3. the TUI views over `stats` and `export` (#455's metrics screen and `s` key), design-gated;
@@ -2,7 +2,7 @@
2
2
 
3
3
  Status: public-safe contributor and agent ramp.
4
4
 
5
- Package/source version in this tree: `0.76.0` (2026-09-03). The Observer is phone-usable as a stated requirement (observer/AGENTS.md); interactive primitives start from Base UI. The Observer renderer is the observer/ workspace artifact only; the legacy string-concat renderer was deleted at cutover (#426), and rollback is a version pin to 0.42.0. The containment boundary introduced in
5
+ Package/source version in this tree: `0.78.0` (2026-09-04). The Observer is phone-usable as a stated requirement (observer/AGENTS.md); interactive primitives start from Base UI. The Observer renderer is the observer/ workspace artifact only; the legacy string-concat renderer was deleted at cutover (#426), and rollback is a version pin to 0.42.0. The containment boundary introduced in
6
6
  `0.15.1` remains in force: managed run and output paths bind to validated
7
7
  physical filesystem identities, and stored provider IDs are evidence, not
8
8
  cleanup authority. The bundled OSS meta-lab is dry-run only until
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "humanish",
3
- "version": "0.76.0",
3
+ "version": "0.78.0",
4
4
  "description": "Open-source-safe CLI for persona simulation, observer review, and public-safe feedback drafts.",
5
5
  "author": "Daniel G Wilson <daniel@danielgwilson.com>",
6
6
  "keywords": [
@@ -103,7 +103,7 @@
103
103
  },
104
104
  "devDependencies": {
105
105
  "@anthropic-ai/claude-agent-sdk": "^0.3.0",
106
- "@e2b/desktop": "^2.2.3",
106
+ "@e2b/desktop": "^2.3.3",
107
107
  "@types/node": "^20.19.41",
108
108
  "@types/pngjs": "^6.0.5",
109
109
  "tsx": "^4.22.4",