gflow-cli 0.9.0__tar.gz → 0.10.0__tar.gz

gflow_cli-0.10.0/.claude/commands/gflow/branch-review.md

@@ -0,0 +1,13 @@
+---
+description: Multi-dimensional LLM council review on the current local feature branch — pre-PR review. Same dimensions as `/gflow:pr-council-review` but reads `git diff <base>..HEAD` instead of a PR; never posts to GitHub. Wrapper around skills/pr-council-review/SKILL.md § 8 (branch mode).
+---
+
+# `/gflow:branch-review [--base <ref>]`
+
+**Read `skills/pr-council-review/SKILL.md` and follow its protocol now in BRANCH MODE (§ 8)**, passing `$ARGUMENTS` (optional `--base <ref>`; default `develop`).
+
+> Do **not** call `Skill(skill="pr-council-review")` — the repo's `skills/*/SKILL.md` files are plain Markdown, not registered as Skill-tool-invocable (only `.claude/commands/gflow/*` are). Invoking it errors with `Unknown skill: pr-council-review`. Read the file directly instead.
+
+The canonical body is `skills/pr-council-review/SKILL.md`. § 8 documents branch-mode-specific behavior: pre-flight, the PR→branch translation table, `release/*` downgrade, SHA drift, and local-only output. All other phases (gather context, detect dimensions, dispatch, synthesize, report) are identical to PR mode.
+
+Sibling: `/gflow:pr-council-review <N>` runs the same council against an open PR.

{gflow_cli-0.9.0 → gflow_cli-0.10.0}/.claude/commands/gflow/check.md

@@ -32,7 +32,7 @@ uv run pyright src
 **4. Tests + coverage** (report only)
 
 ```bash
-uv run pytest -q --cov=gflow_cli --cov-fail-under=80
+uv run python -m pytest -q --cov=gflow_cli --cov-fail-under=80
 ```
 
 ## Output
@@ -46,3 +46,7 @@ uv run pytest -q --cov=gflow_cli --cov-fail-under=80
 
 Ruff fix and format may rewrite multiple files. Always `git diff` before staging.
 Pyright errors and test failures require manual intervention — do not attempt silent workarounds.
+If the coverage run crashes the current MCP/sandbox session with `Connection closed`, re-run the
+same marker-filtered suite in smaller chunks without coverage and rely on CI for the coverage XML.
+Project pytest defaults already exclude `e2e` and `live`; those markers are explicit,
+credit-spending gates and must be requested with a separate `-m e2e` / `-m live` command.

gflow_cli-0.10.0/.claude/commands/gflow/pr-council-review.md

@@ -0,0 +1,11 @@
+---
+description: Multi-dimensional LLM council review of an open PR. Five baseline dimensions (correctness, quality, security, tests, memory-hygiene) plus adaptive per-surface dimensions. Sub-agents invoke specialized skills (security-review, code-review, verify). With no argument, lists open PRs ranked by priority. Wrapper around skills/pr-council-review/SKILL.md (canonical body).
+---
+
+# `/gflow:pr-council-review [PR#]`
+
+**Read `skills/pr-council-review/SKILL.md` and follow its protocol now**, treating `$ARGUMENTS` as the PR number (or empty for prioritize-mode). That file is the canonical body: preflight, dimension detection, parallel dispatch, synthesis, report.
+
+> Do **not** call `Skill(skill="pr-council-review")` — the repo's `skills/*/SKILL.md` files are plain Markdown, not registered as Skill-tool-invocable (only `.claude/commands/gflow/*` are). Invoking it errors with `Unknown skill: pr-council-review`. Read the file directly instead.
+
+Sibling: `/review` is the single-agent Claude-Code built-in (fast, one-pass). Use it for spot-checks; use this command for pre-merge multi-dim audits.

gflow_cli-0.10.0/.claude/commands/gflow/predict.md

@@ -0,0 +1,19 @@
+---
+description: >
+  Pre-implementation 5-persona adversarial analysis for high-stakes gflow-cli proposals.
+  Produces a GO / CAUTION / STOP verdict before any code is written.
+  Invoke before new transports, auth changes, selector redesigns, schema migrations,
+  or any PLAN.md backlog item gated on an investigation step.
+---
+
+# `/gflow:predict [proposal]`
+
+**Read `skills/predict/SKILL.md` and follow its protocol now**, passing `$ARGUMENTS` as the proposal description.
+
+> Do **not** call `Skill(skill="predict")` — the repo's `skills/*/SKILL.md` files are plain Markdown, not registered as Skill-tool-invocable (only `.claude/commands/gflow/*` are). Invoking it errors with `Unknown skill: predict`. Read the file directly instead.
+
+The skill at `skills/predict/SKILL.md` runs five independent expert personas
+(Architect · Security/reCAPTCHA · Performance/Playwright · CLI UX · Devil's Advocate),
+resolves conflicts, and returns a GO / CAUTION / STOP verdict with a confidence score.
+
+**Pair with `/gflow:scenario`** after a GO or CAUTION to enumerate edge cases before EXECUTE.

gflow_cli-0.10.0/.claude/commands/gflow/scenario.md

@@ -0,0 +1,23 @@
+---
+description: >
+  Pre-implementation edge-case explorer for gflow-cli features.
+  Decomposes a proposed change across 12 gflow-cli-specific dimensions
+  (WAF/reCAPTCHA, selector drift, auth lifecycle, batch resume, data layer,
+  cross-platform paths, error propagation, observability) and produces a
+  severity-ranked scenario table and BDD skeleton.
+---
+
+# `/gflow:scenario [feature description]`
+
+**Read `skills/scenario/SKILL.md` and follow its protocol now**, passing `$ARGUMENTS` as the feature or change description.
+
+> Do **not** call `Skill(skill="scenario")` — the repo's `skills/*/SKILL.md` files are plain Markdown, not registered as Skill-tool-invocable (only `.claude/commands/gflow/*` are). Invoking it errors with `Unknown skill: scenario`. Read the file directly instead.
+
+The skill at `skills/scenario/SKILL.md` covers 12 dimensions tuned to gflow-cli's
+known failure surfaces and outputs:
+- A severity-ranked scenario table (Critical / High / Medium / Low)
+- Must-cover acceptance criteria for the PLAN.md task
+- Suggested BDD `Scenario:` blocks for `tests/features/`
+- Cross-references to open KNOWN_ISSUES entries
+
+**Typical order:** `/gflow:predict` → `/gflow:scenario` → PLAN.md task → `/gflow:check` → PR.

{gflow_cli-0.9.0 → gflow_cli-0.10.0}/.env.template

@@ -52,6 +52,34 @@
 # Get one at https://aistudio.google.com/apikey
 # GFLOW_CLI_GEMINI_API_KEY=
 
+# -----------------------------------------------------------------------------
+# Cloud storage
+# -----------------------------------------------------------------------------
+
+# Upload generated assets directly to a cloud bucket instead of saving local
+# asset copies. Supported schemes:
+#   gs://   – Google Cloud Storage (requires: pip install 'gflow-cli[gcs]')
+#   s3://   – Amazon S3 or any S3-compatible store, e.g. MinIO
+#             (requires: pip install 'gflow-cli[s3]')
+#
+# Examples:
+#   GFLOW_CLI_STORAGE_URI=gs://my-gcs-bucket/gflow/
+#   GFLOW_CLI_STORAGE_URI=s3://my-s3-bucket/gflow/
+#   GFLOW_CLI_STORAGE_URI=s3://gflow-test/gflow/   # MinIO local dev
+#
+# For S3 / MinIO you also need the standard AWS SDK env vars:
+#   AWS_ACCESS_KEY_ID=minioadmin
+#   AWS_SECRET_ACCESS_KEY=minioadmin
+#   AWS_ENDPOINT_URL=http://localhost:9000   # omit for real AWS
+#   AWS_DEFAULT_REGION=us-east-1
+#
+# For GCS (local emulator):
+#   STORAGE_EMULATOR_HOST=http://localhost:4443
+#
+# When unset, assets are saved to GFLOW_CLI_OUTPUT_DIR (local disk). gflow-cli
+# does not dual-write local + cloud copies; configure one storage target per run.
+# GFLOW_CLI_STORAGE_URI=
+
 # -----------------------------------------------------------------------------
 # Runtime tuning
 # -----------------------------------------------------------------------------

{gflow_cli-0.9.0 → gflow_cli-0.10.0}/.github/PULL_REQUEST_TEMPLATE.md

@@ -7,6 +7,8 @@
 <!-- List the focused commands you ran, plus any checks you could not run. -->
 
 - [ ] Focused tests added or updated for behavior changes
+- [ ] Documentation updated or explicitly marked not applicable
+- [ ] `uv run python scripts/ci/check_doc_links.py`
 - [ ] `uv run ruff check src tests`
 - [ ] `uv run pyright src`
 - [ ] Relevant pytest command:

{gflow_cli-0.9.0 → gflow_cli-0.10.0}/.github/workflows/ci.yml

@@ -6,7 +6,9 @@ on:
   pull_request:
 
 env:
-  FORCE_JAVASCRIPT_ACTIONS_TO_NODE24: true
+  # Suppress Node.js deprecation warnings and align with user request to use Node 22/remove noise.
+  ACTIONS_ALLOW_USE_UNSECURE_NODE_VERSION: true
+  NODE_VERSION: '22'
 
 jobs:
   # ── Security gate: runs first, on every push/PR, never skippable ──────────
@@ -44,6 +46,9 @@ jobs:
       - name: Repo hygiene (no artefacts, no hardcoded paths)
         run: uv run python scripts/ci/check_repo_hygiene.py
 
+      - name: Documentation links
+        run: uv run python scripts/ci/check_doc_links.py
+
       - name: Lint
         run: uv run ruff check src tests
 
@@ -54,14 +59,10 @@ jobs:
         run: uv run pyright src
 
       - name: Test (smoke only) + coverage
-        # e2e tests (tests/e2e/) require a logged-in Chromium profile
-        # and live Flow auth — never runnable in CI. They self-mark with
-        # `pytestmark = pytest.mark.e2e` (file-level) and are opt-in via
-        # `-m e2e` per their module docstring. `live` marker is reserved for
-        # `GFLOW_LIVE=1` tests that hit the real Flow API. Both are excluded
-        # from the default smoke run; `tests/smoke/test_real_flow.py` already
-        # self-skips via `pytest.mark.skipif(GFLOW_E2E != "1")`.
-        run: uv run python -m pytest -q -m "not e2e and not live" --cov=gflow_cli --cov-report=xml
+        # e2e/live tests require a logged-in Chromium profile and may spend
+        # Flow credits. pyproject.toml excludes both markers by default; CI
+        # intentionally relies on that default smoke set.
+        run: uv run python -m pytest -q --cov=gflow_cli --cov-report=xml
 
       # Upload coverage XML so the sonar job can consume it
       - name: Upload coverage report
@@ -98,7 +99,7 @@ jobs:
         continue-on-error: true  # graceful if coverage upload was skipped
 
       - name: SonarCloud scan
-        uses: SonarSource/sonarcloud-github-action@master
+        uses: SonarSource/sonarcloud-github-action@v5.0.0
         env:
           GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}  # PR decoration
           SONAR_TOKEN: ${{ secrets.SONAR_TOKEN }}    # set in repo → Settings → Secrets

{gflow_cli-0.9.0 → gflow_cli-0.10.0}/.github/workflows/release.yml

@@ -16,7 +16,8 @@ jobs:
       name: pypi
       url: https://pypi.org/p/gflow-cli
     env:
-      FORCE_JAVASCRIPT_ACTIONS_TO_NODE24: true
+      ACTIONS_ALLOW_USE_UNSECURE_NODE_VERSION: true
+      NODE_VERSION: '22'
     steps:
       - uses: actions/checkout@v6
         with:

{gflow_cli-0.9.0 → gflow_cli-0.10.0}/.gitignore

@@ -18,6 +18,7 @@ env/
 pytest-of-*/
 .coverage
 .coverage.*
+coverage.xml
 htmlcov/
 .tox/
 .mypy_cache/
@@ -70,6 +71,8 @@ samples/captured/flow_outgoing_*.jsonl
 samples/captured/flow_outgoing_*.json
 
 # Claude Code worktrees and lock files
+.claude/settings.json
+.claude/settings.local.json
 .claude/worktrees/
 .claude/scheduled_tasks.lock
 
@@ -78,3 +81,5 @@ samples/captured/flow_outgoing_*.json
 
 worktrees/
 
+# E2E test logs (live Flow runs — may contain account/profile names)
+.planning/e2e-logs/

{gflow_cli-0.9.0 → gflow_cli-0.10.0}/AGENTS.md

@@ -2,7 +2,7 @@
 
 > Universal entry point for AI coding agents. Read this first; everything else routes from here.
 
-Supported tools that auto-discover this file: Cursor, Codex, Aider, Gemini CLI, Jules, Devin, Windsurf, Zed, Warp, opencode, RooCode, Amp, Junie, Phoenix, GitHub Copilot, VS Code, Factory, Augment, Semgrep, Kilo Code, UiPath. Claude Code reads [CLAUDE.md](CLAUDE.md), which cross-references this file.
+Supported tools that auto-discover this file: Cursor, Codex, Aider, Jules, Devin, Windsurf, Zed, Warp, opencode, RooCode, Amp, Junie, Phoenix, GitHub Copilot, VS Code, Factory, Augment, Semgrep, Kilo Code, UiPath. Gemini CLI reads [GEMINI.md](GEMINI.md) and Claude Code reads [CLAUDE.md](CLAUDE.md), both of which cross-reference this file.
 
 ## Project at a glance
 
@@ -37,6 +37,7 @@ Run these gates in order before every commit:
 ```powershell
 $env:PYTHONUTF8=1
 uv run python scripts/ci/check_repo_hygiene.py
+uv run python scripts/ci/check_doc_links.py
 uv run ruff check src tests
 uv run ruff format --check src tests
 uv run pyright src
@@ -45,8 +46,9 @@ uv run python -m pytest -q --cov=gflow_cli
 
 Or invoke the wrapper: `/gflow:check`.
 
-- Use `pytest -m "not live and not e2e"` locally; full suite OOMs on small dev machines. Scope to changed dirs; trust CI for the full sweep.
+- Use `pytest -m "not live and not e2e and not smoke"` locally; full suite OOMs on small dev machines. Scope to changed dirs; trust CI for the full sweep.
 - TDD is non-negotiable. Coverage floor: 80% overall.
+- Documentation is a first-class deliverable. Every behavior, workflow, config, or operator-facing change must update the relevant docs or state why no docs changed in the PR/checklist. `scripts/ci/check_doc_links.py` is a merge gate.
 - Live tests (`@pytest.mark.live`) opt in via `GFLOW_LIVE=1`. E2E tests require `GFLOW_CLI_E2E_PROFILE`.
 
 ## Code style

{gflow_cli-0.9.0 → gflow_cli-0.10.0}/CHANGELOG.md

@@ -7,6 +7,270 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [0.10.0] — 2026-05-29
+
+### Fixed
+
+- **Image and video counts leaking across profiles in `gflow data list projects` (issue #113).**
+  Fixed an issue where project media counts were combined if two different profiles
+  happened to share the same `flow_project_id`. The counts in `_LIST_PROJECTS_SQL`
+  are now strictly scoped to the active `profile_name`.
+
+- **List queries fan-out when multiple operations claim an asset (issue #111).**
+  Fixed a bug in `gflow data list images` and `gflow data list videos` where 
+  assets would be duplicated or have non-deterministic prompts if multiple 
+  operations (e.g., retries) claimed the same output asset. The SQL queries 
+  now use a deterministic subquery grouping by `asset_id` to ensure exactly 
+  one-to-one cardinality.
+
+### Added
+
+- **Google account identity persisted to every profile (`issue #92`).** Both
+  auth strategies (`real_chrome`, `internal_chromium`) now write a
+  `.gflow_account` file to the profile directory immediately after the session
+  is verified, durably associating the signed-in email with the profile on disk.
+  `ProfileMeta` gains a `google_account: str | None` field populated by
+  `profile_store.list_profiles()` from that file. `gflow auth list` (table and
+  `--json`) now includes a **Google account** column so every profile is
+  immediately identifiable — no more opaque `default` entries.  The `--json`
+  output gains the `google_account` key for programmatic callers.  Closes #92.
+
+- **Auto-rename of the first-run `default` profile to email local-part.** When
+  `gflow auth login` creates the first profile and no `--profile` flag was given,
+  the profile is named `default` as a placeholder. After the session is verified
+  and the email is known, `auth login` automatically renames `profile_default` to
+  `profile_<email-local-part>` (e.g. `profile_ffroliva`) and updates
+  `config.toml`'s `default_profile` pointer atomically. The local-part is
+  sanitized to a filesystem-safe name (characters outside letters, digits, `-`,
+  and `_` become `-`), so `flavio.oliva@gmail.com` → `profile_flavio-oliva` and
+  `user+flow@gmail.com` → `profile_user-flow`. Existing `default`
+  profiles that were created before this change continue to work; they gain the
+  email column the next time `gflow auth login` is run against them.  Closes #92.
+
+- **`profile_store.rename_profile(old_name, new_name)`** — reusable primitive that
+  renames a profile directory and updates `config.toml` when the renamed profile
+  was the default. Raises `FileNotFoundError` / `FileExistsError` on invalid input.
+
+- **Zero-credit smoke test for profile account persistence
+  (`tests/smoke/test_profile_account_smoke.py`).** Three smoke tests that verify
+  the full observable chain — account file present + readable, `list_profiles()`
+  surfaces `google_account`, `gflow auth list --json` includes the field — against
+  a real authenticated profile. No image or video generation; zero Flow credits.
+  Backfills the `.gflow_account` file for profiles created before the fix so the
+  tests work on existing sessions. Opt-in via
+  `GFLOW_CLI_E2E_PROFILE=<name> pytest -m smoke tests/smoke/test_profile_account_smoke.py`.
+
+- **Aggregated asset view in `gflow data list images/videos`.** By default,
+  listing images or videos now returns one row per asset (Flow media ID),
+  collapsing multiple local copies into a single entry with a `COPIES` count
+  and the path of the latest copy. This prevents duplicate rows when
+  re-downloading the same media to different directories. The `copy_count`
+  is also exposed in JSONL output.
+- **`--all-copies` flag on `gflow data list images/videos`.** Restores the
+  previous behavior of showing every local file as a separate row.
+- **`gflow data prune` command.** New maintenance utility to remove stale
+  `local_files` database entries for local paths that no longer exist on
+  disk. Only targets local files (ignores cloud-stored assets). Supports
+  `--dry-run` to preview deletions and `--profile` to limit the scan.
+- **External storage documentation for S3, MinIO, and Google Cloud Storage.**
+  Adds `docs/EXTERNAL_STORAGE.md`, cross-links it from the README, docs index,
+  configuration, data-layer, usage, security, and user-guide docs, and clarifies
+  that `GFLOW_CLI_STORAGE_URI` is a cloud-only output mode rather than
+  local-plus-cloud dual-write.
+- **`--json` flag on `gflow video t2v`, `gflow video i2v`, and `gflow video r2v`.**
+  Emits the `VideoResult` (status / command / media_id / generation_status /
+  succeeded / local_path / failure_reasons / error_message) plus the request
+  echo (model / mode / aspect / duration / count / seed) as a single JSON
+  object on stdout. A failed generation still emits its JSON payload and
+  then exits 1. The data-layer recorder
+  (`record_started_video` / `record_completed_video`) fires regardless of
+  `--json` so audit history is independent of the output channel. E2e
+  coverage for `--json` shape across image / video / auth / models lives at
+  `tests/e2e/test_json_output_e2e.py`.
+- **`--json` flag on `gflow image t2i` and `gflow image i2i`.** Emits the
+  complete `GeneratedImage` result (every field — `media_name`,
+  `workflow_id`, `seed`, `prompt`, `model_name_type`, `aspect_ratio`,
+  `dimensions`, `fife_url`, `is_signed_url`) plus the on-disk `local_path`
+  as a single JSON object on stdout. A worker keys `images[0].seed` for
+  refine-regen seed continuity. Single-prompt only (`--json` rejects
+  multi-prompt batch with a Click usage error); progress chatter is
+  suppressed so stdout is pure JSON. `ref_count` surfaces only on i2i.
+- **`gflow models` catalog command.** New top-level command that enumerates
+  the image and video model catalog as a Rich table (default) or as a single
+  JSON object (`--json`). Per-model: `name`, CLI aliases (filtered to what
+  the generation command's `--model` Choice actually accepts), `ref_cap`,
+  `default` (image), `max_duration` (video). Built from the
+  `Model` / `VideoModel` enums + their alias maps so the catalog can never
+  drift from what the generation commands accept. A UI populating its model
+  picker from `gflow models --json` is guaranteed to pass any selected alias
+  back to `--model`.
+- **`gflow_cli.json_output` module + `--json` error path on
+  `run_with_handlers`.** Pure builders for image/video result payloads and
+  RFC 9457 problem-details errors (plus a `retryable` flag worker schedulers
+  key their retry-vs-absorb decision off — WAF / rate-limit / network /
+  timeout). When `as_json=True`, errors emit a parseable JSON payload on
+  stdout with the same exit code as the Rich path; the observability event
+  still fires.
+- **`--json` flag on `gflow auth list`.** Emits the profile inventory as a JSON
+  array (`name` / `is_default` / `cookies_present` / `profile_dir` /
+  `last_used_at`) on stdout so a programmatic caller (e.g. a worker discovering
+  authenticated profiles) can `json.loads(stdout)` instead of scraping the
+  Rich table.
+- **Per-model r2v reference-image cap rebuilt as a data table.** Replaces the
+  prior pair of constants (`OMNI_REFERENCE_CAP=7`, `VEO_REFERENCE_CAP=3`) with
+  a `VideoModel -> int` mapping consulted by
+  `gflow_cli.api.video.reference_cap_for(model)`. New entries:
+  `veo_3_1_lite_lower_priority=3` (was implicitly covered by the veo branch),
+  and `veo_3_1_quality=0` — Veo 3.1 Quality does NOT support
+  Ingredients/References to Video at all per Google Flow's official support
+  page; passing it to r2v raises a clear `does not support R2V
+  (reference-to-video)` error rather than letting the request fail at the
+  wire. CLI guard added on `gflow video r2v` (`click.UsageError`, exit 2)
+  mirroring the i2i pattern so over-cap and quality+r2v fail before any
+  profile/network work. E2e tripwire at
+  `tests/e2e/test_video_r2v_ref_cap_e2e.py` asserts Flow actually consumes
+  all `cap` refs at the at-cap boundary.
+- **Per-model i2i reference-image cap.** Flow silently keeps only the first N
+  reference images when an i2i request attaches more than the model accepts,
+  so a caller could believe every ref was used. `gflow_cli.api.image.reference_cap_for(model)`
+  exposes the live-observed per-model cap: NARWHAL (Nano Banana 2) and GEM_PIX_2
+  (Nano Pro) accept 10, IMAGEN_3_5 (Imagen 4) accepts 3. Enforced as a domain
+  invariant in `GenerateImageRequest.__post_init__` and at the CLI boundary in
+  `gflow image i2i` (clean `click.UsageError` / exit 2 before any
+  profile/network work). Mirrors the existing video r2v cap pattern. E2e
+  tripwire at `tests/e2e/test_image_i2i_ref_cap_e2e.py` asserts Flow actually
+  consumes all `cap` refs (one `reference_attached` event per ref) so a future
+  silent truncation on the Flow side fails the test.
+- **Layered e2e test strategy with cost sub-markers.** The single `e2e` marker is
+  now augmented by cost sub-markers (`e2e_auth`, `e2e_image`, `e2e_video`,
+  `e2e_batch`, `e2e_data`, `smoke`) so callers can run only the tier they can
+  afford (zero-credit auth checks, single-credit image smoke, etc.). See
+  `docs/E2E_TESTING.md` for the full reference.
+- `tests/e2e/conftest.py` — shared `e2e_profile_dir`, `e2e_nosession_profile`,
+  and `e2e_env` fixtures replace duplicated inline helpers in individual test files.
+- `tests/api/transports/test_transport_timeout.py` — extracted from e2e as a
+  pure-mock integration test; also fixes `Path("/dev/null")` → `Path(os.devnull)`
+  for Windows portability.
+- `tests/test_marker_registry.py` — invariant checks that every e2e test carries a
+  cost sub-marker, and self-tests for the auto-marker conftest hook.
+- `docs/E2E_TESTING.md` — comprehensive e2e strategy and layer reference document.
+
+### Fixed
+
+- Structlog logs are now routed to stderr (via
+  `PrintLoggerFactory(file=sys.stderr)`) instead of stdout. Previously every
+  CLI event leaked onto stdout, which broke the `--json` contract for
+  programmatic callers — `json.loads(stdout)` failed because the JSON
+  payload was preceded by event-log lines. Logs are diagnostics; stdout is
+  data. No-op for human users (terminals still show logs the same way).
+
+### Changed
+
+- `gflow data media` now labels cloud-backed asset records as `cloud_uri_N`
+  while keeping local assets under the existing `local_path_N` labels.
+- `gflow_cli.api.video` no longer exposes the standalone `OMNI_REFERENCE_CAP` /
+  `VEO_REFERENCE_CAP` constants. Callers that need a per-model R2V cap should
+  use `reference_cap_for(model)` (which returns `0` for `VEO_3_1_QUALITY` —
+  R2V is unsupported there). `MAX_REFERENCE_IMAGES` (= 7) is unchanged and
+  still the absolute ceiling used when the model is unknown.
+- `GFLOW_CLI_E2E_RUN_VIDEO` default flipped from `"1"` to `"0"`. The Veo step in
+  `test_data_layer_e2e.py` is now **opt-in**: set `GFLOW_CLI_E2E_RUN_VIDEO=1` to
+  include it. This prevents accidental Veo credit burns on unattended CI runs.
+- `pytest` `addopts` now excludes `smoke` in addition to `e2e` and `live`:
+  `not e2e and not live and not smoke`. Bare `pytest` never launches a live
+  browser session.
+- Auto-marker conftest hook uses `item.path.parts` instead of a slash-delimited
+  string substring, fixing Windows backslash path compatibility.
+
+## [0.9.1] — 2026-05-27
+
+> **Locale and catalog patch release.** Hardens the headed-browser UI
+> automation path for localized Flow profiles, fixes I2V start/end-frame
+> attachment on non-English Chrome/Flow sessions, and repairs first-run catalog
+> edge cases found after v0.9.0.
+
+### Changed
+
+- `NEW_PROJECT_SELECTORS` now covers all 14 supported locales (EN / PT / ES /
+  FR / DE / IT / NL / JA / ZH / KO / PL / RU / TR / ID) and leads with
+  locale-stable icon selectors (`add_2` Material Symbols ligature on
+  `<button>` and on `[role='button']` ARIA-role variants, plus an anchored
+  `^\+\s+\S+$` regex for `+ <word>` host elements). English-only
+  `[aria-label*='New project']` and `[aria-label*='Project']` ARIA fallbacks
+  removed.
+  `SUBMIT_BUTTON_SELECTORS` drops its English-only
+  `button[aria-label*="Create"]` fallback — the preceding `arrow_forward` icon
+  entries already cover this button in every locale. Both selector tuples are
+  now fully locale-invariant for non-English Chrome profiles. The `--lang=en-US`
+  Chromium launch arg is retained only to stabilise `IMAGE_MODEL_OPTION_SELECTORS`
+  (English product names); its removal is tracked as issue #24 Phase 5 (#94).
+
+### Fixed
+
+- Bare `pytest` no longer collects live/e2e tests by default. The project-wide
+  pytest `addopts` now excludes `e2e` and `live` unless callers explicitly pass
+  a different `-m` expression, and `tests/smoke/test_real_flow.py` is marked
+  with both markers so marker-filtered local and CI runs cannot accidentally
+  launch a real Flow browser session.
+- Browser-manager PID tests no longer call the real `os.kill` while pretending
+  to be on POSIX from a Windows runner. The POSIX liveness branches are now
+  tested with mocked `os.kill`, avoiding hard interpreter/session exits during
+  local test runs.
+
+- Running the pytest suite no longer writes fixture rows into the developer's
+  production `gflow.db` catalog. A new autouse `_isolate_settings` fixture in
+  `tests/conftest.py` redirects `GFLOW_CLI_HOME` and `GFLOW_CLI_DB_PATH` to
+  per-test `tmp_path` dirs and clears the `get_settings()` `lru_cache` before
+  and after every test, preventing the cached singleton from ever resolving to
+  a `platformdirs` production path. Closes
+  [#86](https://github.com/ffroliva/gflow-cli/issues/86).
+
+- `gflow video i2v` no longer silently breaks on non-English Chrome profiles.
+  PR #70's structural-first `_attach_frame` cascade matched **zero** real
+  slots — its anchor selector assumed the `swap_horiz` icon used class
+  `google-symbols` (it uses `material-icons`) and the slots were `<button>`
+  (they're `<div type="button">`). Production I2V therefore relied on the
+  English-text fallback, which silently misses on any non-EN profile (pt-BR
+  shows `Inicial`/`Final`, DE shows `Anfang`/`Ende`, etc.). Replaced
+  `FRAME_SLOTS_STRUCT` with the locale-free pattern
+  `div[type='button'][aria-haspopup='dialog']` and added a `.first`-of-remaining
+  fallback for the End-frame case (after Start is attached, only one slot
+  matches and the prior `.nth(slot_index)` went out-of-bounds). Live-verified
+  with `tests/e2e/test_transports_e2e.py::test_e2e_i2v_start_end_frame_attach`
+  on `ffroliva` + `GFLOW_CLI_LOCALE=de-DE` (Chrome rendered pt-BR; both
+  non-EN). Closes [#63](https://github.com/ffroliva/gflow-cli/issues/63).
+
+### Changed
+
+- `gflow data media <id>` now searches across **all** profiles by default,
+  matching the cross-profile default of `gflow data list`. Pass
+  `--profile NAME` to disambiguate the rare case where the same Flow
+  media ID exists under multiple profiles (the command refuses to
+  guess and prints the list of candidate profiles, each annotated with
+  its `kind`). Closes
+  [#87](https://github.com/ffroliva/gflow-cli/issues/87).
+
+### Fixed
+
+- `gflow data list` no longer crashes with `no such table: assets` on a
+  missing or freshly-created catalog DB. The query path now routes through
+  `DataStore.open`, which applies schema migrations on first connect —
+  first-time users and anyone recovering from a wiped DB get an empty
+  table and exit 0 instead of a `DataStoreError`. Closes
+  [#88](https://github.com/ffroliva/gflow-cli/issues/88).
+- `gflow auth list` no longer crashes with `UnicodeEncodeError` on Windows
+  consoles whose code page cannot encode the default-profile marker `●`
+  (cp1252 in PowerShell / cmd by default). The renderer now picks a glyph
+  safe for the active `sys.stdout.encoding` — `●` on UTF-8, ASCII `*` on
+  cp1252 / ascii / latin-1 / unknown. Closes [#82](https://github.com/ffroliva/gflow-cli/issues/82).
+
+### Documentation
+
+- `PLAN.md` refreshed to reflect develop state through v0.9.0 — marks Phase 6
+  (data layer) shipped via PR #58 + #78 + #81, Phase 7 Issue #24 Phase 2
+  shipped via PR #70, Phase B I2V/R2V shipped via PR #48, and resolves the
+  duplicate Phase 7 numbering (pluggable storage renumbered to Phase 8).
+
 ## [0.9.0] — 2026-05-25
 
 > **Maturity & Visibility release.** Surfaces the SQLite catalog (PR #52/#58)
@@ -945,7 +1209,9 @@ shell-script template that branches on these codes.
 
 First skeleton. Not functional end-to-end yet.
 
-[Unreleased]: https://github.com/ffroliva/gflow-cli/compare/v0.9.0...HEAD
+[Unreleased]: https://github.com/ffroliva/gflow-cli/compare/v0.10.0...HEAD
+[0.10.0]: https://github.com/ffroliva/gflow-cli/compare/v0.9.1...v0.10.0
+[0.9.1]: https://github.com/ffroliva/gflow-cli/compare/v0.9.0...v0.9.1
 [0.9.0]: https://github.com/ffroliva/gflow-cli/compare/v0.8.1...v0.9.0
 [0.8.1]: https://github.com/ffroliva/gflow-cli/compare/v0.8.0...v0.8.1
 [0.8.0]: https://github.com/ffroliva/gflow-cli/compare/v0.7.0...v0.8.0

{gflow_cli-0.9.0 → gflow_cli-0.10.0}/CONTRIBUTING.md

@@ -39,6 +39,11 @@ def test_parse_uuid_from_url(): ...
 async def test_upload_returns_asset(): ...
 
 @pytest.mark.e2e               # Hits the real Flow API. Requires GFLOW_CLI_E2E_PROFILE env var.
+@pytest.mark.e2e_image         # Cost sub-marker: spends ~1 Imagen credit.
+async def test_full_t2i_roundtrip(): ...
+
+@pytest.mark.e2e
+@pytest.mark.e2e_video         # Cost sub-marker: spends ~1 Veo credit (most expensive).
 async def test_full_i2v_roundtrip(): ...
 ```
 
@@ -46,10 +51,23 @@ CI runs `unit` + `integration` on every push. `e2e` tests require a live authent
 
 ```bash
 export GFLOW_CLI_E2E_PROFILE=<profile-name>   # name of a logged-in profile
-uv run pytest -m e2e -v
+
+# Zero-credit sanity check (auth + health)
+uv run pytest -m e2e_auth -v
+
+# Single image (1 Imagen credit)
+uv run pytest -m "e2e_image and not e2e_batch" -v
+
+# Full regression (all credits)
+GFLOW_CLI_E2E_RUN_VIDEO=1 uv run pytest -m e2e -v
 ```
 
-E2e tests spend real Veo/Imagen credits — run them on `develop` before opening a release PR to `main`. See [docs/DEVELOPMENT.md § E2e gate](docs/DEVELOPMENT.md#e2e-gate-before-merging-develop--main).
+E2e tests spend real Veo/Imagen credits. Video tests default to opt-out — set
+`GFLOW_CLI_E2E_RUN_VIDEO=1` to include them. Run the full suite on `develop`
+before opening a release PR to `main`.
+
+See [docs/E2E_TESTING.md](docs/E2E_TESTING.md) for the complete layer reference,
+cost table, and run commands.
 
 ### Coverage targets
 
@@ -64,13 +82,19 @@ E2e tests spend real Veo/Imagen credits — run them on `develop` before opening
 
 ```bash
 uv run python scripts/ci/check_repo_hygiene.py  # artefact + path hygiene
+uv run python scripts/ci/check_doc_links.py     # internal Markdown links
 uv run ruff check src tests          # lint
 uv run ruff format src tests         # auto-format
 uv run pyright src                   # type-check (strict on src/gflow_cli/)
 uv run pytest -q --cov=gflow_cli      # tests + coverage
 ```
 
-CI runs all five on every push. Install local pre-commit hooks (recommended):
+CI runs all six on every push. Documentation is part of the merge gate: update
+the relevant docs for behavior, workflow, configuration, architecture, or
+operator-facing changes. If no documentation change is needed, state that in
+the PR validation checklist.
+
+Install local pre-commit hooks (recommended):
 
 ```bash
 pip install pre-commit && pre-commit install

gflow_cli-0.10.0/GEMINI.md

@@ -0,0 +1,27 @@
+# GEMINI.md
+
+> Project memory hub for **Gemini CLI**. The universal coding-agent rules for any tool (Cursor, Codex, Aider, Gemini CLI, etc.) live in [AGENTS.md](AGENTS.md) — this file carries Gemini-specific session protocol only.
+
+## What this project is
+
+`gflow-cli` is an unofficial Python CLI that drives [Google Flow](https://labs.google/fx/tools/flow) (Veo image-to-video, Imagen text-to-image) from the terminal by reverse-engineering Flow's private REST API. See [README.md](README.md) for the user-facing overview.
+
+## On every session start
+
+1. Read **[AGENTS.md](AGENTS.md)** — universal rules every agent must follow.
+2. Read **[docs/INDEX.md](docs/INDEX.md)** — routing layer for all project docs and commands.
+3. Pull deeper context on demand:
+   - Starting a feature → `/gflow:plan`
+   - Touching auth or reCAPTCHA → `/gflow:known-issues`
+   - Cutting a release → `/gflow:release`
+   - Before any commit → `/gflow:check`
+
+## Gemini-specific
+
+- Use specialized skills when relevant (e.g., `find-docs` for library research, `pr-council-review` for PR audits).
+- Maintain memory via the `mcp-mempalace` tool if available.
+- Prioritize **turn efficiency** and **high-signal output**.
+
+## Active phase
+
+See [PLAN.md](PLAN.md) or run `/gflow:plan` for the current detailed plan.