Prepare Python package release 0.3.2

CHANGELOG.md

@@ -4,6 +4,22 @@ All notable changes are documented here. FSQ follows Semantic Versioning after p
 
 ## [Unreleased]
 
+## [0.3.2] - 2026-10-21
+
+Release preparation; this entry does not establish a published package, tag, or GitHub release. See the [release notes](docs/releases/v0.3.2.md) for upgrade steps and validation limits.
+
+### Fixed
+
+- Android text-entry verification tolerates geometry changes only when a unique locator remains bound to the original resource, while retaining exact readback and target-identity checks. Ambiguous, unreadable, disabled, or mismatched targets remain non-success results without implicit write retries (#39).
+- Android's initial device-context read retries once for the specifically identified `ApplicationSharedMemory not initialized` backend error; it does not repeat the application action or retry unrelated failures (#39).
+
+### Changed
+
+- Refreshed the bilingual README, Landing Page, current Android Quick Test demonstration, captions, screenshots, and media source records. The Website retains its headline and white/green visual identity, activates inline native video only on user action, and preserves independent viewing links (#41).
+- Both public entry paths now lead to Quick Test through Coding Agent or FSQ UI as equal alternatives, with one complete repository-search → About website → homepage-verification example, evidence review, separate advanced Case/replay workflows, and Star/feedback destinations. Documentation contract tests follow the current onboarding (#41).
+- Removed the duplicate YouTube cover-image links from both READMEs while retaining the GitHub inline video, text links, and three explanatory images (#42).
+- Python distribution metadata is prepared for version 0.3.2. The private npm package remains independently versioned at 0.1.0; this metadata preparation adds no runtime or dependency change.
+
 ## [0.3.1] - 2026-10-20
 
 Release preparation; this entry does not establish a published package, tag, or GitHub release. See the [release notes](docs/releases/v0.3.1.md) for upgrade steps and validation limits.
@@ -73,8 +89,11 @@ Release preparation; this entry does not establish a published package, tag, or
 
 - This is an alpha release. Public APIs and Case authoring details may evolve before 1.0.
 
-[Unreleased]: https://github.com/FSQ-lab/FSQ/compare/v0.3.1...HEAD
+[Unreleased]: https://github.com/FSQ-lab/FSQ/compare/v0.3.2...HEAD
+[0.3.2]: https://github.com/FSQ-lab/FSQ/compare/v0.3.1...v0.3.2
 [0.3.1]: https://github.com/FSQ-lab/FSQ/compare/v0.3.0...v0.3.1
 [0.1.6]: https://github.com/FSQ-lab/FSQ/compare/v0.1.5...v0.1.6
 [0.1.5]: https://github.com/FSQ-lab/FSQ/compare/v0.1.0...v0.1.5
 [0.1.0]: https://github.com/FSQ-lab/FSQ/releases/tag/v0.1.0
+
+The 0.3.2 comparison references anticipate release tags; they do not establish that a tag or publication exists.

docs/releases/v0.3.2.md

@@ -0,0 +1,112 @@
+# FSQ v0.3.2 release preparation
+
+Date: 2026-10-21. Python distribution: `fsq-agent` 0.3.2. FSQ remains alpha.
+This document prepares a release identity and notes; it does not establish a
+`v0.3.2` tag, GitHub release, PyPI publication, or deployed Website revision.
+
+## Changes since the 0.3.1 preparation
+
+- [Android input and context fixes (#39)](https://github.com/FSQ-lab/FSQ/pull/39):
+  text-entry readback can recognize the same uniquely resource-bound control
+  after keyboard or layout changes. It retains exact text, package, class,
+  resource, editability, and uniqueness checks. Positional locators do not
+  gain permission to follow a different input. Masked/unreadable values,
+  ambiguous or disabled targets, mismatches, and lost identity cannot produce
+  success; no alternative target or implicit write retry is introduced.
+  Initial device context also retries one read for the specifically identified
+  `ApplicationSharedMemory not initialized` backend failure. Other errors
+  propagate normally; no application action is repeated by that recovery.
+- [Public presentation and onboarding (#41)](https://github.com/FSQ-lab/FSQ/pull/41):
+  synchronized English/Chinese README and guides with the Landing Page, while
+  preserving the Website headline and white/green visual system. Coding Agent
+  and FSQ UI are equal Quick Test entry paths with expandable setup, visible
+  prerequisites, and private Provider/model setup. The shared example searches
+  for FSQ-lab/FSQ, follows its About website link, and verifies both homepage
+  headlines. Case management and deterministic replay remain separately chosen
+  advanced workflows. Star and sanitized-feedback links follow the trial.
+  The current Android demonstration, matching explanatory screenshots, captions,
+  cover, social image, and source records replace the active old presentation.
+  Native Website playback starts only after activation; controls, captions,
+  keyboard focus, and fallback viewing links remain available.
+- [README demo simplification (#42)](https://github.com/FSQ-lab/FSQ/pull/42):
+  removed the duplicate YouTube cover-image links from both READMEs. The GitHub
+  inline video, YouTube/caption/provenance text links, and three story images
+  remain. Website playback and its cover asset are unchanged by that trim.
+- Python metadata and its exact distribution-contract assertion advance to
+  0.3.2. The private npm package retains its independent 0.1.0 version.
+  This release-preparation edit does not change runtime code, dependencies,
+  public interfaces, or configuration.
+
+These changes are established by repository history following the merged
+0.3.1 preparation (#38). The comparison references below anticipate release
+tags, rather than claiming an earlier package or tag was published.
+
+## Upgrade after publication
+
+Once 0.3.2 is published and obtainable from the package index used by your
+Python environment, upgrade with:
+
+```bash
+python -m pip install --upgrade fsq-agent==0.3.2
+```
+
+Project-local Skills are copies. Package installation or a host reload alone
+does not refresh them. From your Coding Agent project, explicitly run:
+
+```bash
+fsq init --agent skill
+```
+
+Review installer updates, backups, and conflicts under `.agents/skills/fsq/`,
+then reload Skill discovery or start a new Coding Agent session.
+`--no-overwrite` preserves an existing copy. No Workspace reinitialization or
+Provider reconfiguration is required solely for this version change.
+
+Continue through either [Coding Agent or FSQ UI](../getting-started.md).
+Quick Test uses the dedicated user-level Workspace, returns inspectable
+results/evidence, and does not save a formal Case or automatically replay one.
+Python 3.11+, a configured supported Provider and selected model, and
+[platform-specific prerequisites](../platform-prerequisites.md) still apply.
+Never paste Provider credentials into Coding Agent chat.
+
+## Validation and publication status
+
+The preparation uses a failing exact-version contract check before the metadata
+bump, followed by the focused distribution-contract module, Ruff lint/format,
+and diff checks. Results belong to the final identified source snapshot and its
+external validation evidence; these instructions do not assert a passing result.
+Earlier PR, release, or local results are historical and are not final 0.3.2
+validation evidence.
+
+Final frontend compilation, wheel and sdist creation, a wheel rebuilt from the
+sdist, metadata/resource comparisons, fresh source-free installed-package smoke
+tests, and independent review are pending integration verification. The smoke
+must use the built artifact outside the source checkout, including both console
+entry points, platform presets, copied Skill resources, and packaged static
+serving. These checks do not replace a real target or Provider test.
+
+The complete 0.3.2 Python/frontend CI and supported-platform matrix are pending.
+No all-platform live execution, clean public-index end-to-end first-use run, or
+successful release pipeline is claimed. Building a local wheel does not prove
+public installation availability.
+
+[Media provenance](../media/quick-test-demo-provenance.md) records the matching
+Android Run's producer as FSQ 0.2.1; it is not a recording of 0.3.2. Source commit,
+capture/export timestamps, and some runtime conditions remain unestablished.
+The [media acceptance record](../media/media-acceptance.md) separately retains
+whole-video privacy/security, complete caption synchronization and muted
+comprehension, three-person new-user acceptance, public first-use, and deployed
+canonical-page verification limits. Selected-image review is not full-video
+privacy clearance. Evidence may contain private application data; review it
+before sharing. These publication checks are not satisfied by a version bump,
+unit test, or build.
+
+No tag creation, workflow dispatch, GitHub release, PyPI upload, or deployment
+is performed by this preparation. Publication requires a separate authorized
+release action and its existing manual verification/environment prerequisites.
+
+- [Anticipated v0.3.1 → v0.3.2 comparison](https://github.com/FSQ-lab/FSQ/compare/v0.3.1...v0.3.2)
+- [Unreleased comparison after v0.3.2](https://github.com/FSQ-lab/FSQ/compare/v0.3.2...HEAD)
+
+These links use the canonical repository and do not establish that the tags or
+published artifacts exist.

pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "fsq-agent"
-version = "0.3.1"
+version = "0.3.2"
 description = "Goal-driven FSQ automated testing agent with a local agent engine, OpenAI-compatible model access, harness actions, local utilities, observation, and reporting."
 readme = "README.md"
 license = "MIT"

tests/test_distribution_contract.py

@@ -79,7 +79,7 @@ def test_default_distribution_contract() -> None:
     metadata = project["project"]
 
     assert metadata["name"] == "fsq-agent"
-    assert metadata["version"] == "0.3.1"
+    assert metadata["version"] == "0.3.2"
     version = Version(metadata["version"])
     assert version.local is None
     dependencies = metadata["dependencies"]