gflow-cli 0.41.0__tar.gz → 0.42.0__tar.gz

{gflow_cli-0.41.0 → gflow_cli-0.42.0}/.claude/commands/gflow/llm-council.md

@@ -1,10 +1,10 @@
 ---
-description: Run pr-council-review (PR or branch mode) plus an external-tools corroboration layer (codex, gemini, opt-in agy) for high-stakes reviews where a same-model-family blind spot is a real risk. Tiers: small (internal only, default) / medium (+codex) / high (+codex+gemini). Wrapper around skills/llm-council/SKILL.md.
+description: Run pr-council-review (PR or branch mode) plus an external-tools corroboration layer (codex, plus Antigravity via the agy harness) for high-stakes reviews where a same-model-family blind spot is a real risk. Tiers: small (internal only, default) / medium (+codex) / high (+codex+antigravity). Wrapper around skills/llm-council/SKILL.md.
 ---
 
-# `/gflow:llm-council [PR# | --base <ref>] [--tier small|medium|high] [--include-agy]`
+# `/gflow:llm-council [PR# | --base <ref>] [--tier small|medium|high]`
 
-**Read `skills/llm-council/SKILL.md` and follow its protocol now**, treating `$ARGUMENTS` as: the PR number or `--base <ref>` (branch mode, same as `/gflow:pr-council-review` / `/gflow:branch-review`), plus `--tier <small|medium|high>` (default `small`) and the optional `--include-agy` flag.
+**Read `skills/llm-council/SKILL.md` and follow its protocol now**, treating `$ARGUMENTS` as: the PR number or `--base <ref>` (branch mode, same as `/gflow:pr-council-review` / `/gflow:branch-review`), plus `--tier <small|medium|high>` (default `small`). The `high` tier adds Antigravity (`agy`) alongside `codex`.
 
 > Do **not** call `Skill(skill="llm-council")` or `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 either errors with `Unknown skill: ...`. Read the files directly instead.
 

{gflow_cli-0.41.0 → gflow_cli-0.42.0}/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, 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.
+Supported tools that auto-discover this file: Cursor, Codex, Aider, Antigravity (`agy`), 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.
 
 ## Project at a glance
 
@@ -92,7 +92,7 @@ The `skills/` directory ships installable agent skill docs in plain Markdown wit
 | `plan` | [`skills/plan/SKILL.md`](skills/plan/SKILL.md) | Create a structured task-by-task implementation plan for a feature |
 | `status` | [`skills/status/SKILL.md`](skills/status/SKILL.md) | Show current plan state, progress, and next unchecked task |
 | `pr-council-review` | [`skills/pr-council-review/SKILL.md`](skills/pr-council-review/SKILL.md) | Multi-dimensional PR council review |
-| `llm-council` | [`skills/llm-council/SKILL.md`](skills/llm-council/SKILL.md) | Wraps `pr-council-review` with external CLI reviewers (`codex`/`gemini`, opt-in `agy`) for high-stakes reviews |
+| `llm-council` | [`skills/llm-council/SKILL.md`](skills/llm-council/SKILL.md) | Wraps `pr-council-review` with external CLI reviewers (`codex` + Antigravity `agy`) for high-stakes reviews |
 | `issue-assessment` | [`skills/issue-assessment/SKILL.md`](skills/issue-assessment/SKILL.md) | Triage a GitHub issue (read-only) before any fix work |
 | `issue-resolve` | [`skills/issue-resolve/SKILL.md`](skills/issue-resolve/SKILL.md) | Drive an assessed issue to a test-first fix + draft PR |
 | `check` | [`skills/check/SKILL.md`](skills/check/SKILL.md) | Quality gates (lint/format/types/tests) before every commit |
@@ -108,7 +108,7 @@ and must be resolvable by ANY agent via this file. Vendor directories hold thin
 only (`.claude/commands/gflow/*.md` are pointers into `skills/`); never put protocol
 content in a vendor directory.
 
-**Cursor / Aider / Codex / Gemini CLI (`agy`):** paste or include the relevant `SKILL.md` in your system context. Note: in the `agy` TUI prompt, custom slash commands (e.g. `/gflow:pr-council-review`) are blocked by the TUI's command parser. Type them as plain text without the leading slash (e.g. `gflow:pr-council-review`, `gflow:branch-review`, or `gflow:check`) to trigger the corresponding agent skill workflow.
+**Cursor / Aider / Codex / Antigravity (`agy`):** paste or include the relevant `SKILL.md` in your system context. Note: in the `agy` TUI prompt, custom slash commands (e.g. `/gflow:pr-council-review`) are blocked by the TUI's command parser. Type them as plain text without the leading slash (e.g. `gflow:pr-council-review`, `gflow:branch-review`, or `gflow:check`) to trigger the corresponding agent skill workflow.
 **Claude Code:** the `/gflow:` slash commands in `.claude/commands/gflow/` are auto-discovered when the project is open — no extra setup needed. To register a skill globally, copy the command file to `~/.claude/commands/`.
 **Custom agents:** fetch `skills/gflow-cli/SKILL.md` into your knowledge base before answering gflow questions.
 

{gflow_cli-0.41.0 → gflow_cli-0.42.0}/CHANGELOG.md

@@ -7,6 +7,62 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [0.42.0] — 2026-07-21
+
+### Fixed
+
+- Content-safety `400` responses from Flow are now classified as
+  `ContentPolicyError` (exit code 5) instead of the misleading `WireFormatError`,
+  so callers can branch on a policy rejection deterministically. (#359)
+- Corrected a false `flow_operation_id` invariant. Live verification found
+  `veo-lite` can produce a `remote_started` checkpoint with `operation_id` unset,
+  contradicting the prior claim that only `omni-flash` omits the operation name.
+  `flow_operation_id` is best-effort/optional; `media_id` is the canonical handle
+  used by polling, download, and every CLI lookup. No behaviour change; a
+  regression test now locks the operation-name parser against the real veo
+  capture fixture. (#361)
+
+### Changed
+
+- **Replaced Gemini CLI with Antigravity (`agy`) as the supported Google coding
+  agent** across skills and docs (Google retired Gemini CLI in favor of the
+  Antigravity harness). Antigravity auto-discovers `AGENTS.md` natively, so the
+  dedicated `GEMINI.md` hub is **removed** and Antigravity is listed among the
+  auto-discovering tools in `AGENTS.md`. (#360)
+- **`llm-council` skill:** the `high` tier's second external reviewer is now
+  Antigravity (`agy`) instead of Gemini CLI; `agy` is promoted from an opt-in
+  tool to the pinned `high`-tier slot and the `--include-agy` flag is dropped.
+  When `agy` is unavailable or fails non-interactively, the skill suggests
+  installing Antigravity or substituting another external CLI coding agent rather
+  than failing. (#360)
+- Forward-looking agent references updated for consistency across `AGENTS.md`,
+  `CLAUDE.md`, `README.md`, `llms.txt`, `docs/INDEX.md`, `docs/AGENT_GUIDE.md`,
+  `skills/*`, and CI helper strings. Historical release notes and past
+  verification records are left unchanged. (#360)
+
+### Removed
+
+- Removed the dead `fail_processing_tasks` daemon-recovery method, superseded by
+  the checkpoint-classifying `recover_processing` (internal cleanup; no
+  user-facing behaviour change). (#361)
+
+### Documentation
+
+- Added an onboarding page and quickstart flow to the published docs site, and
+  anonymized personal account data across it. Security vulnerabilities now report
+  through GitHub's private vulnerability reporting rather than a personal email. (#362)
+- Corrected the same-profile concurrency framing (MCP page and docs) to the
+  cross-process `ProfileLease` fail-fast contract (`ProfileLockedError`, exit 11),
+  and recorded that the POSIX `fcntl.flock` lease was verified green on the
+  Windows/macOS/Linux CI matrix. (#361)
+
+### Tests
+
+- Added live-gated end-to-end tests for crash-recovery (a crashed post-submit task
+  recovers as `indeterminate` and is never resubmitted) and cancellation-safe
+  teardown (the profile lease releases on mid-launch cancel). Both were confirmed
+  against real Flow on 2026-07-21. (#361)
+
 ## [0.41.0] — 2026-07-20
 
 ### Added
@@ -2169,7 +2225,8 @@ shell-script template that branches on these codes.
 
 First skeleton. Not functional end-to-end yet.
 
-[Unreleased]: https://github.com/ffroliva/gflow-cli/compare/v0.41.0...HEAD
+[Unreleased]: https://github.com/ffroliva/gflow-cli/compare/v0.42.0...HEAD
+[0.42.0]: https://github.com/ffroliva/gflow-cli/compare/v0.41.0...v0.42.0
 [0.41.0]: https://github.com/ffroliva/gflow-cli/compare/v0.40.0...v0.41.0
 [0.40.0]: https://github.com/ffroliva/gflow-cli/compare/v0.39.0...v0.40.0
 [0.39.0]: https://github.com/ffroliva/gflow-cli/compare/v0.38.1...v0.39.0

{gflow_cli-0.41.0 → gflow_cli-0.42.0}/CLAUDE.md

@@ -1,6 +1,6 @@
 # CLAUDE.md
 
-> Project memory hub for **Claude Code**. The universal coding-agent rules for any tool (Cursor, Codex, Aider, Gemini CLI, etc.) live in [AGENTS.md](AGENTS.md) — this file carries Claude-Code-specific session protocol only.
+> Project memory hub for **Claude Code**. The universal coding-agent rules for any tool (Cursor, Codex, Aider, Antigravity, etc.) live in [AGENTS.md](AGENTS.md) — this file carries Claude-Code-specific session protocol only.
 
 ## What this project is
 

{gflow_cli-0.41.0 → gflow_cli-0.42.0}/KNOWN_ISSUES.md

@@ -585,31 +585,37 @@ one of the 14 supported locales or in English.
 
 ---
 
-### `omni-flash` t2v response omits operation name — `flow_operation_id` persists NULL
+### t2v response can omit operation name — `flow_operation_id` persists NULL
 
 - **Status:** Open · **Severity:** Low · **Affects:** v0.9.0+ (data layer)
 
 The data layer's `on_started` callback captures
 `operations[0].operation.name` from each `batchAsyncGenerateVideoText`
-response and persists it as `operations.flow_operation_id`. The
-`omni-flash` model's response shape does not carry that field, so
-omni-flash rows end up with `flow_operation_id` NULL while `veo-*` rows
-carry the expected `operations/...` identifier. The rest of the row
-(prompt, model, aspect, started/completed timestamps, batch ID, output
-paths) is recorded normally.
-
-**Impact:** cosmetic for now — `gflow data media <id>` and provenance
-lookup by Flow media ID still work. Any future feature that joins on
-`flow_operation_id` (none in the current CLI) would miss omni-flash
-rows.
+response and persists it as `operations.flow_operation_id`. This field
+is **best-effort, not guaranteed for any model**: `omni-flash`'s response
+shape does not carry it, so omni-flash rows end up with
+`flow_operation_id` NULL. **Update 2026-07-21 (live):** the same NULL
+was observed on a live `veo-lite` t2v run (`remote_started` checkpoint
+with `operation_id=None`), on an agentic-UI-cohort account — so `veo-*`
+is not a reliable guarantee either; the original claim that veo-* rows
+always carry the operation name was false. The rest of the row (prompt,
+model, aspect, started/completed timestamps, batch ID, output paths) is
+recorded normally regardless.
+
+**Impact:** cosmetic/provenance-only — `media_id` is the canonical
+handle (poll, download, and every CLI lookup use it, not
+`flow_operation_id`). `gflow data media <id>` and provenance lookup by
+Flow media ID still work. Nothing in the current CLI queries by
+`flow_operation_id`; a future feature that did would miss rows on any
+model whose response omits the operation name.
 
 **Workaround:** none needed if you don't query by `flow_operation_id`.
 
-**Roadmap:** capture an `omni-flash` `batchAsyncGenerateVideoText`
-response sample, identify the equivalent provenance handle (if any), and
-either map it into `flow_operation_id` or document that omni-flash
-legitimately has no such identifier. Track via a follow-up issue once a
-sample is captured.
+**Roadmap:** capture response samples across models (including veo-lite)
+where the operation name is absent, identify the equivalent provenance
+handle (if any), and either map it into `flow_operation_id` or document
+that these cases legitimately have no such identifier. Track via a
+follow-up issue once samples are captured.
 
 ---
 

{gflow_cli-0.41.0 → gflow_cli-0.42.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: gflow-cli
-Version: 0.41.0
+Version: 0.42.0
 Summary: Unofficial CLI for Google Flow — drive Veo image-to-video generations from the terminal.
 Project-URL: Homepage, https://github.com/ffroliva/gflow-cli
 Project-URL: Issues, https://github.com/ffroliva/gflow-cli/issues
@@ -161,7 +161,7 @@ gflow-cli ships four agent entry points. Pick the one your tool reads first.
 
 | File | Audience | Tools |
 |---|---|---|
-| [**AGENTS.md**](AGENTS.md) | Universal coding-agent spec | Cursor · Codex · Aider · Gemini CLI · Jules · Devin · Windsurf · Zed · Warp · opencode · Copilot |
+| [**AGENTS.md**](AGENTS.md) | Universal coding-agent spec | Cursor · Codex · Aider · Antigravity · Jules · Devin · Windsurf · Zed · Warp · opencode · Copilot |
 | [**CLAUDE.md**](CLAUDE.md) | Claude Code's auto-loaded memory | Claude Code |
 | [**llms.txt**](llms.txt) | LLM-readable summary (llmstxt.org format) | Paste into ChatGPT, Claude, or Gemini to onboard the model |
 | [`skills/gflow-cli/SKILL.md`](skills/gflow-cli/SKILL.md) | Claude Code Skill | Symlink into `~/.claude/skills/` |

{gflow_cli-0.41.0 → gflow_cli-0.42.0}/PLAN.md

@@ -369,7 +369,7 @@ The video-generation feature has its own sub-phase plan (spike → Phase A → P
 - [ ] **Integrated Filmmaking Studio, Flow Worker & MCP SSE Service.** (Backlog) The MCP→FlowWorker wiring shipped in v0.23.0 (PR #228). Remaining: a local FastAPI/React filmmaking studio dashboard + HTTP/SSE MCP. Planned in [gflow-studio-scaffold/PLAN.md](file:///C:/development/github/gflow-cli/docs/superpowers/plans/2026-06-24-gflow-studio-scaffold/PLAN.md) and [rest-api-layer/PLAN.md](file:///C:/development/github/gflow-cli/docs/superpowers/plans/2026-06-24-rest-api-layer/PLAN.md).
 - [x] **Issue #92: Google account identity persistence shipped 2026-05-28 (PR #110).** After every `gflow auth login`, both auth strategies write the verified email to `$GFLOW_CLI_HOME/profile_<name>/.gflow_account`. `ProfileMeta.google_account` surfaces it; `gflow auth list` and `gflow auth list --json` display it. First-login auto-rename: when no `--profile` flag was given and the only profile is named `default`, it is renamed to the email local-part. `profile_store.rename_profile()` primitive added. Zero-credit smoke test (`tests/smoke/test_profile_account_smoke.py`) ships with backfill path for pre-existing profiles. LLM council (9-dim) run; all must-fix items applied before merge. See `KNOWN_ISSUES.md` § Resolved for full detail.
 - [x] **`/gflow:pr-council-review` slash command shipped 2026-05-26 (PR #97).** Multi-dimensional LLM council for open PRs (4 baseline + 8 adaptive dimensions, mandatory memory-slug binding, draft-PR guard, YELLOW escape valve, live-verify gate). Validated on PR #93; self-audited by 3-agent meta-council surfacing 13 must-fix items applied before merge. See memory `[[llm-council-code-review-pr93]]`.
-  - **Phase A (Backlog — Portability):** extract command body into `skills/pr-council-review/SKILL.md` so Gemini CLI / Codex / Cursor / Aider can consume it. Keep `.claude/commands/gflow/pr-council-review.md` as a thin wrapper. Memory: `[[pr-council-review-portability-backlog]]`.
+  - **Phase A (Backlog — Portability):** extract command body into `skills/pr-council-review/SKILL.md` so Antigravity / Codex / Cursor / Aider can consume it. Keep `.claude/commands/gflow/pr-council-review.md` as a thin wrapper. Memory: `[[pr-council-review-portability-backlog]]`.
   - **Phase B (Backlog — Token optimization):** add `scripts/dev/pr_council_prefetch.py` (single gh-call producing structured JSON for agents) and `scripts/dev/memory_filter.py` (relevant slug bodies only). Targets ~30-50% per-agent context cost reduction.
   - **Phase C (Backlog — Reusable meta-council):** codify the 3-dim audit (completeness / robustness / prompt-clarity) we used on this command as `/gflow:meta-council-audit <path>` for re-use on future skills/commands.
   - Sequencing: **A → C → B**.

{gflow_cli-0.41.0 → gflow_cli-0.42.0}/README.md

@@ -90,7 +90,7 @@ gflow-cli ships four agent entry points. Pick the one your tool reads first.
 
 | File | Audience | Tools |
 |---|---|---|
-| [**AGENTS.md**](AGENTS.md) | Universal coding-agent spec | Cursor · Codex · Aider · Gemini CLI · Jules · Devin · Windsurf · Zed · Warp · opencode · Copilot |
+| [**AGENTS.md**](AGENTS.md) | Universal coding-agent spec | Cursor · Codex · Aider · Antigravity · Jules · Devin · Windsurf · Zed · Warp · opencode · Copilot |
 | [**CLAUDE.md**](CLAUDE.md) | Claude Code's auto-loaded memory | Claude Code |
 | [**llms.txt**](llms.txt) | LLM-readable summary (llmstxt.org format) | Paste into ChatGPT, Claude, or Gemini to onboard the model |
 | [`skills/gflow-cli/SKILL.md`](skills/gflow-cli/SKILL.md) | Claude Code Skill | Symlink into `~/.claude/skills/` |

{gflow_cli-0.41.0 → gflow_cli-0.42.0}/docs/AGENT_GUIDE.md

@@ -124,7 +124,7 @@ is reserved in `check_materiality.py` (which already classifies every change).
 ### Satisfying a gate without Claude Code
 
 Every gate enforces a **deliverable**, not a Claude-specific command. Non-Claude agents
-(Cursor / Codex / Gemini CLI / Aider) and humans read the relevant `skills/<name>/SKILL.md`
+(Cursor / Codex / Antigravity / Aider) and humans read the relevant `skills/<name>/SKILL.md`
 directly and produce the same artifact — e.g. read `skills/predict/SKILL.md` and write the
 5-persona verdict into the PR, instead of running `/gflow:predict`.
 

{gflow_cli-0.41.0 → gflow_cli-0.42.0}/docs/INDEX.md

@@ -6,7 +6,7 @@ Welcome to the `gflow-cli` documentation. This index is the routing layer: it te
 |---|---|---|
 | [README](../README.md) | Project overview, install, quick start | First time landing on the repo |
 | [docs/DEMOS.md](DEMOS.md) | Gallery of `gflow` in action (terminal + split-screen Flow recordings) | You want to see what gflow looks like running |
-| [AGENTS.md](../AGENTS.md) | Universal coding-agent spec — Cursor / Codex / Aider / Gemini CLI / Jules / etc. | Any AI coding agent enters the repo |
+| [AGENTS.md](../AGENTS.md) | Universal coding-agent spec — Cursor / Codex / Aider / Antigravity / Jules / etc. | Any AI coding agent enters the repo |
 | [llms.txt](../llms.txt) | LLM-readable summary (llmstxt.org format) | A user pastes context about gflow into ChatGPT / Claude / Gemini |
 | [docs/PROJECT_STATUS.md](PROJECT_STATUS.md) | Full milestone history + lifecycle policy | Auditing where the project is in its lifecycle |
 | [docs/AGENT_GUIDE.md](AGENT_GUIDE.md) | Mandates and routing rules for AI agents (companion to AGENTS.md) | A coding agent needs the longer non-negotiable rules |
@@ -46,7 +46,7 @@ Welcome to the `gflow-cli` documentation. This index is the routing layer: it te
 | **[docs/AGENT_UI_RECON.md](AGENT_UI_RECON.md)** | Agentic Flow UI cohort: classic-vs-agentic DOM signature, server-side + volatile (flapping) A/B gating (no client-readable flag), why `_exit_agent_mode` fails, runtime-DOM-detection recommendation | Touching `_exit_agent_mode` / composer mode-switch, diagnosing exit 23 selector drift, or understanding Flow's agentic-UI A/B (#183/#174) |
 | **[docs/ASSET_TAGGING_RECON.md](ASSET_TAGGING_RECON.md)** | `@`-mention asset tagging research & integration design: what Flow's `@CharacterName`/`@me`/`@ingredient` tags are, how they map onto the already-verified `referenceEntities`/`referenceImages` wire, the H1-vs-H2 serialization hypothesis, the CLI-side mention-resolver proposal, and the pre-implementation capture spike plan | Working on inline `@` mentions in prompts, `video --character` reuse, or planning the mention-capture spike |
 | **[tasks/lessons.md](../tasks/lessons.md)** | Running notebook of patterns + reviewer findings, dated and traced to commits | Starting a new phase; debugging "why did the council flag this?" |
-| **[skills/README.md](../skills/README.md)** | Installable agent skill docs (gflow-cli, predict, pr-council-review, scenario) — cross-tool portable Markdown consumed by Claude Code, Cursor, Codex, Gemini CLI, Aider, etc. | Any agent wanting to use gflow-cli correctly |
+| **[skills/README.md](../skills/README.md)** | Installable agent skill docs (gflow-cli, predict, pr-council-review, scenario) — cross-tool portable Markdown consumed by Claude Code, Cursor, Codex, Antigravity, Aider, etc. | Any agent wanting to use gflow-cli correctly |
 | **[scripts/dev/skillopt/README.md](../scripts/dev/skillopt/README.md)** | SkillOpt mock harness — rollout→score loop for measuring and improving skill doc accuracy across multiple LLM providers | Measuring a skill edit's impact; comparing Claude vs GPT-4o vs Gemini on gflow tasks |
 | **[scripts/diag/README.md](../scripts/diag/README.md)** | Diagnostic investigation scripts — run against a live authenticated profile to capture wire samples, measure Chrome memory, or mint reCAPTCHA tokens | Running a one-off investigation against a live Flow session; establishing baseline measurements for issue #155 |
 | [docs/superpowers/specs/2026-07-04-pr-triage-autopilot-design.md](superpowers/specs/2026-07-04-pr-triage-autopilot-design.md) | PR-triage autopilot design spec: Stage 0/1 gate, Docker sandboxing, Telegram and audit ledger | Reviewing or auditing the automated PR-triage setup |
@@ -74,7 +74,7 @@ Slash commands for Claude Code, stored in `.claude/commands/gflow/`. All prefixe
 | `/gflow:scenario <feature>` | 12-dimension edge-case explorer → severity-ranked scenario table + BDD skeleton | After predict GO/CAUTION; before `/gflow:plan` |
 | `/gflow:pr-council-review [PR#]` | Multi-dimensional council review of an open PR (5 baseline + adaptive dimensions) | Before merging any non-trivial PR; mandatory for auth/transport/data changes |
 | `/gflow:branch-review` | Same council review run against the current local feature branch (no PR needed) | Pre-PR self-audit; after predict/scenario on a high-stakes branch |
-| `/gflow:llm-council [PR#\|--base <ref>] [--tier small\|medium\|high]` | Wraps `pr-council-review` with external CLI reviewers (`codex`/`gemini`, opt-in `agy`) for cross-model corroboration | High-stakes reviews where a same-model-family blind spot is a real risk |
+| `/gflow:llm-council [PR#\|--base <ref>] [--tier small\|medium\|high]` | Wraps `pr-council-review` with external CLI reviewers (`codex` + Antigravity `agy`) for cross-model corroboration | High-stakes reviews where a same-model-family blind spot is a real risk |
 | `/gflow:doc-review` | Systematic council-driven audit of documentation completeness and drift | Before cutting a release; after major documentation changes |
 
 **Governance:** commands are executable docs — they decay like any doc. When a phase advances or a file path changes, update the relevant command in the same commit. `/gflow:release` includes a staleness review step.
@@ -114,7 +114,7 @@ Slash commands for Claude Code, stored in `.claude/commands/gflow/`. All prefixe
 **"Where can I look up a media ID I generated yesterday?"** → [DATA_LAYER § Querying the data layer](DATA_LAYER.md#querying-the-data-layer)
 **"How do I stop gflow from storing my prompts?"** → [DATA_LAYER § Privacy and redaction](DATA_LAYER.md#privacy-and-redaction)
 **"What does exit code 16 mean and how do I recover?"** → [DATA_LAYER § Persistence-failure handling](DATA_LAYER.md#persistence-failure-handling)
-**"How do I use this project's skills in Cursor / Codex / Gemini CLI / Aider?"** → [skills/README.md](../skills/README.md#use-with-other-agents)
+**"How do I use this project's skills in Cursor / Codex / Antigravity / Aider?"** → [skills/README.md](../skills/README.md#use-with-other-agents)
 **"How do I benchmark a skill doc against real tasks?"** → [scripts/dev/skillopt/README.md](../scripts/dev/skillopt/README.md)
 **"How do I compare Claude vs GPT-4o vs Gemini on gflow tasks?"** → `python scripts/dev/skillopt/harness.py --provider openai --model gpt-4o` (see [skillopt README](../scripts/dev/skillopt/README.md))
 **"What does Chrome actually use in RAM during a generation?"** → run `uv run python scripts/diag/memory_profile.py --profile NAME` (see [scripts/diag/README.md](../scripts/diag/README.md))
@@ -132,7 +132,7 @@ Slash commands for Claude Code, stored in `.claude/commands/gflow/`. All prefixe
 **"A gflow command hangs / fails — where do I start?"** → [DEBUGGING § Quick reference](DEBUGGING.md#quick-reference)
 **"Flow's UI broke a selector — how do I diagnose it?"** → [DEBUGGING § Inspecting Flow's live UI](DEBUGGING.md#inspecting-flows-live-ui)
 **"What does each `ui_automation.*` log event mean?"** → [DEBUGGING § Listener & HTTP-layer debugging](DEBUGGING.md#listener--http-layer-debugging)
-**What was actually live-verified for the latest release?** → latest: [LIVE_VERIFICATION_v0.40.0](LIVE_VERIFICATION_v0.40.0.md) (**#344 `@`-mention asset-tagging resolution proven live e2e, ~2 Imagen credits** — `gflow character create` staged a real taggable character [bound reference images, `has_reference_images` guard], the e2e test's documented propagation wait cleared Flow's `projectInitialData` lag, and `gflow image t2i "@Zoro …"` passed with returncode 0, a real generated image, and the correctly de-tagged prompt in the catalog — no code changes needed, `resolve_and_apply` as merged in #344 already works end-to-end; pre-existing unrelated issue #174 [Flow media-library A/B flake] can still intermittently block generation regardless of this feature — recorded, not omitted). Prior: [v0.39.0](LIVE_VERIFICATION_v0.39.0.md) (**#341 failed-generation persistence proven live** — a real `image t2i` on denon82 failed at the wire [`batchGenerateImages` HTTP 400 → `WireFormatError`] and the record-on-catch funnel wrote the **first-ever `failed` row** into the production catalog [`error_type=wire-format`, redacted `error_detail`, browsable via `gflow data list errors`], **$0**; the remaining paid paths [video/chain/batch/movie/worker] are funnel-regression-tested, not force-failed live [would cost credits], and the status-aware `succeeded=false` video-poll fix is unit-locked — recorded, not omitted). Prior: [v0.38.1](LIVE_VERIFICATION_v0.38.1.md) (**#338 agentic-pin recovery proven live ON the pinned account** — run 1 caught the composer-render race the unit suite could not [`ensure_media_incomplete` +115 ms, toggle never reached]; the readiness-wait fix landed and run 2 recovered classic on denon82 after ~2h of pin [`ui_driver.bound mode=classic`, real 768×1376 JPEG, exact prompt adherence]; reload branch unit-locked, live trigger not observed this cycle — recorded, not omitted). Prior: [v0.38.0](LIVE_VERIFICATION_v0.38.0.md) (**#332 mode control + #314 i2i ref dedup proven live e2e** — `image t2i --ui-mode classic` recovered agentic→classic in a real run [`ui_mode.attempt_exit_agent` → full classic chain, 1376×768 JPEG]; two identical `image i2i` runs proved the dedup contract [run 1: picker miss → `image_uploaded`; run 2: `reference_deduped_by_filename`, ZERO uploads]; `video t2v` veo-fast [`MEDIA_GENERATION_STATUS_SUCCESSFUL`, 15.2 MB `ftyp` mp4]; autopilot #333 recorded deferred-with-reason [hermes-ops deployment staged separately]). Prior: [v0.37.0](LIVE_VERIFICATION_v0.37.0.md) (**viewport 1920×1080 + agentic count enforcement + FIPS-safe SAPISIDHASH proven live e2e** — `image t2i` [classic cohort, real 1376×768 JPEG] and `video t2v` veo-fast [`MEDIA_GENERATION_STATUS_SUCCESSFUL`, 11.5 MB `ftyp` mp4] both succeeded end-to-end at the new 1920×1080 viewport with every selector resolving [no drift]; SAPISIDHASH digest verified byte-identical [FIPS-safe `usedforsecurity=False`], auth unchanged). Prior: [v0.36.0](LIVE_VERIFICATION_v0.36.0.md) (**#316 diagnostic tooling live-verified against real Flow traffic** — `GFLOW_CLI_HAR_PATH` captured a real 54-entry HAR [real Flow domains, 34× 200s, and a genuine 401 the tool exists to diagnose]; `GFLOW_CLI_DEBUG_TRACEBACK`'s typed-error scoping confirmed live [console output identical with/without the flag on a `GFlowError`]; unhandled-path behavior covered by 6 new tests (30 total across both files) + SDD task review; #312 reference-entity-smuggling fix covered by its own dedicated e2e test). Prior: [v0.35.0](LIVE_VERIFICATION_v0.35.0.md) (**#305-follow-up multimodal reverse-engineering & storyboard tool proven live e2e** — `gflow tools run storyboard` successfully deconstructed a cyberpunk scene into a cohesive 4-panel sequence; `reverse-engineer` tool extracts frames from video streams/URLs using `claude-video`'s `watch.py` and produces structured prompt components using Gemini multimodal models; dynamic token budgeting and strict Pyright types verified via test suite). Prior: [v0.34.0](LIVE_VERIFICATION_v0.34.0.md) (**#299 `--ui-mode` gate proven live, ZERO credits** — drove `get_ui_driver` directly over 4 real editor loads, never generated: bidirectional switch works [classic→agentic via `_force_agent_mode`, agentic→classic via `_exit_agent_mode`, each DOM-reverified], `agentic`-when-already-agentic is idempotent, and a timed-out toggle → **exit-28 fail-fast before submission**; deprecation behavior-change [`prefer_classic`→exit 28] confirmed). Prior: [v0.33.0](LIVE_VERIFICATION_v0.33.0.md) (**#287-follow-up picker fix proven live on the original repro** — the v0.32.1-failing `i2v <uuid> --project <crowded>` command now exits 0 with a real 3.0 MB `ftyp` mp4, `frame_ref_attached` + `video_saved`, zero uploads; **#241 jitter live-verified on both paths** — flag `--jitter 4-6` → `batch_jitter_sleep seconds=4.98`, and the new 0.5–1.5 s default → `seconds=1.47`, each firing exactly once between two t2i submissions; environment note: the denon82 cohort is now natively agentic and round-2 generations hit the pre-existing #281 `MediaAttributionError`, not a regression). Prior: [v0.32.1](LIVE_VERIFICATION_v0.32.1.md) (**#293 teardown fix live-verified credit-free** — error-path aborts leave ZERO chrome.exe holding the profile [3× leak reproduced pre-fix same day]; force-close fallback + ProfileLockedError translation unit-pinned; #283 wave logic fixes mock-verified to exact call counts). Prior: [v0.32.0](LIVE_VERIFICATION_v0.32.0.md) (**#287 UUID i2v frames proven live e2e** — an in-project asset selected by media UUID in the frame-slot dialog [frame_ref_attached, startImage bound on the wire, ZERO upload events] produced a real 720×1280 interpolation, resolving the #237 negative prior; foreign UUID → exit 9 pre-generation with screenshot; **#288 duration fail-fast proven live** — exit 23 pre-submit saved a 10-credit generation, and the debug screenshot confirmed the duration control is ABSENT from the cohort's settings popover [locale hypothesis refuted]; upload-400 exit-27 repro deferred-with-reason — the original rejected JPEG no longer exists). Prior: [v0.31.0](LIVE_VERIFICATION_v0.31.0.md) (**#281 wrong-media defenses live-verified in BOTH directions, credit-free** — correct attribution downloads with real wire metadata [seed, 768×1376]; the incident-class lazy-render ambiguity reproduced live and hard-stopped with `MediaAttributionError` exit 26 naming every candidate UUID, no file written; **#282 two-UUID-ref i2i proven e2e** — both freshly generated portrait UUIDs picker-selected into one scene with both canonical identities visually transferred). Prior: [v0.30.0](LIVE_VERIFICATION_v0.30.0.md) (**agentic-cohort image path + both fixes live-verified credit-free** — native 768×1376 stills, i2i-chain + canonical-ref identity hold; the character-create `DataIntegrityError` and the `prefer_classic` WARNING were observed live then fixed [recorder idempotent + INFO log]; MCP video-param feature deferred-with-reason on Veo credits, per the stills-only goal). Prior: [v0.29.0](LIVE_VERIFICATION_v0.29.0.md) (**persistent `gflow instructions` CRUD proven live e2e GREEN, credit-free** — 11-command sequence on a fresh project: add incl. `--ref <image-UUID>` attach, list/enable/disable, master toggle both ways, TOML `apply` destructive full-sync with `enabled=false` honored, `rm` with server card-id preserved across mutations; movie brief-sync code-path-verified via the same primitives, deferred-with-reason on credits; MCP `gflow_instructions_*` tools are thin adapters over the live-verified primitives). Prior: [v0.28.0](LIVE_VERIFICATION_v0.28.0.md) (**agent instructions `-i` now steer agentic generation — crayon e2e GREEN**: a style-neutral prompt + an enabled "crayon" card produced an unmistakable crayon drawing through the real transport; root causes were conversational phrasing + the `project_brief.enabled` master switch). Prior: [v0.27.1](LIVE_VERIFICATION_v0.27.1.md) (**v0.27.0 release follow-up fixes and documentation sync**). Prior: [v0.27.0](LIVE_VERIFICATION_v0.27.0.md) (**`[style]` block with named variants + prompt-aware resume — credit-free CLI verification**: real `movie run --dry-run` 3-way ledger proving resolved-style display, `skip (done)` on matching `style_hash`, and `re-run (style changed)` after a style edit; composition is local, wire path untouched, so no Veo credits spent — first credited styled run will confirm visually). Prior: [v0.26.0](LIVE_VERIFICATION_v0.26.0.md) (**image i2i by generated-image UUID proven live e2e GREEN** — the existing asset is *selected in place* by its media UUID, no duplicate upload, producing an i2i output that references it; plus generated-image `display_name` capture, credited @C1ph3r404). Prior: [v0.25.0](LIVE_VERIFICATION_v0.25.0.md) (**#237 remote-UUID i2v proven live e2e GREEN** — a real 8s 720×1280 interpolation from a generated image's UUID, after the picker-search attach was reworked to local upload; #240 home-`.env` matrix live; plus two silent-failure guards root-caused live: video-as-image download rejection and rejected-upload fail-loud). Prior: [v0.24.0](LIVE_VERIFICATION_v0.24.0.md) (--project parity across CLI + MCP — automated coverage; live gen auth-gated). Prior: [v0.23.0](LIVE_VERIFICATION_v0.23.0.md) (MCP generation wiring proven live end-to-end — tool→worker→real Flow REST→structured-error translation; final image-write blocked by an expired session, environmental. #222 macOS fix reporter-verified e2e on Apple Silicon). Prior: [v0.22.0](LIVE_VERIFICATION_v0.22.0.md) (Tools framework — automated/CI coverage complete; credit/key-gated live feature run is a pending owner gate, see the doc) · [v0.21.0](LIVE_VERIFICATION_v0.21.0.md) (MCP server over stdio + HTTP/SSE — credit-free protocol handshake; 4 tools / 2 prompts / 3 resources confirmed on stdout) · [v0.20.1](LIVE_VERIFICATION_v0.20.1.md) · [v0.20.0](LIVE_VERIFICATION_v0.20.0.md) · [v0.19.0](LIVE_VERIFICATION_v0.19.0.md) · [v0.18.0](LIVE_VERIFICATION_v0.18.0.md) · [v0.17.0](LIVE_VERIFICATION_v0.17.0.md) · [v0.16.0](LIVE_VERIFICATION_v0.16.0.md) · [v0.13.0](LIVE_VERIFICATION_v0.13.0.md) · [v0.12.0](LIVE_VERIFICATION_v0.12.0.md) · [v0.11.0](LIVE_VERIFICATION_v0.11.0.md) · [v0.10.0](LIVE_VERIFICATION_v0.10.0.md) · [v0.9.1](LIVE_VERIFICATION_v0.9.1.md) · [v0.9.0](LIVE_VERIFICATION_v0.9.0.md) · [v0.8.1](LIVE_VERIFICATION_v0.8.1.md) · [v0.7.0](LIVE_VERIFICATION_v0.7.0.md)
+**What was actually live-verified for the latest release?** → latest: [LIVE_VERIFICATION_v0.42.0](LIVE_VERIFICATION_v0.42.0.md) (**#361 production-readiness live-gaps confirmed live 2026-07-21, 1 Veo credit** — a real `veo-lite` t2v through the worker path persisted a real `remote_started` checkpoint [observer→checkpoint→DB chain confirmed], crash-recovery classified the task `indeterminate` without resubmitting, and mid-launch cancellation released the `ProfileLease` credit-free; full evidence in [v0.40.0-production-readiness § "2026-07-21 live follow-up"](LIVE_VERIFICATION_v0.40.0-production-readiness.md#2026-07-21-live-follow-up-0420-cycle); #359 content-safety `400`→`ContentPolicyError` classification is unit-covered, not live-triggered [would require a deliberate policy violation]; #360/#362 are docs-only, verified by the documentation gates). Prior: [v0.41.0](LIVE_VERIFICATION_v0.41.0.md) (**#357 production-readiness hardening — stale-session fail-fast, free image gen, paid veo-lite T2V all proven live**). Prior: [v0.40.0](LIVE_VERIFICATION_v0.40.0.md) (**#344 `@`-mention asset-tagging resolution proven live e2e, ~2 Imagen credits** — `gflow character create` staged a real taggable character [bound reference images, `has_reference_images` guard], the e2e test's documented propagation wait cleared Flow's `projectInitialData` lag, and `gflow image t2i "@Zoro …"` passed with returncode 0, a real generated image, and the correctly de-tagged prompt in the catalog — no code changes needed, `resolve_and_apply` as merged in #344 already works end-to-end; pre-existing unrelated issue #174 [Flow media-library A/B flake] can still intermittently block generation regardless of this feature — recorded, not omitted). Prior: [v0.39.0](LIVE_VERIFICATION_v0.39.0.md) (**#341 failed-generation persistence proven live** — a real `image t2i` on denon82 failed at the wire [`batchGenerateImages` HTTP 400 → `WireFormatError`] and the record-on-catch funnel wrote the **first-ever `failed` row** into the production catalog [`error_type=wire-format`, redacted `error_detail`, browsable via `gflow data list errors`], **$0**; the remaining paid paths [video/chain/batch/movie/worker] are funnel-regression-tested, not force-failed live [would cost credits], and the status-aware `succeeded=false` video-poll fix is unit-locked — recorded, not omitted). Prior: [v0.38.1](LIVE_VERIFICATION_v0.38.1.md) (**#338 agentic-pin recovery proven live ON the pinned account** — run 1 caught the composer-render race the unit suite could not [`ensure_media_incomplete` +115 ms, toggle never reached]; the readiness-wait fix landed and run 2 recovered classic on denon82 after ~2h of pin [`ui_driver.bound mode=classic`, real 768×1376 JPEG, exact prompt adherence]; reload branch unit-locked, live trigger not observed this cycle — recorded, not omitted). Prior: [v0.38.0](LIVE_VERIFICATION_v0.38.0.md) (**#332 mode control + #314 i2i ref dedup proven live e2e** — `image t2i --ui-mode classic` recovered agentic→classic in a real run [`ui_mode.attempt_exit_agent` → full classic chain, 1376×768 JPEG]; two identical `image i2i` runs proved the dedup contract [run 1: picker miss → `image_uploaded`; run 2: `reference_deduped_by_filename`, ZERO uploads]; `video t2v` veo-fast [`MEDIA_GENERATION_STATUS_SUCCESSFUL`, 15.2 MB `ftyp` mp4]; autopilot #333 recorded deferred-with-reason [hermes-ops deployment staged separately]). Prior: [v0.37.0](LIVE_VERIFICATION_v0.37.0.md) (**viewport 1920×1080 + agentic count enforcement + FIPS-safe SAPISIDHASH proven live e2e** — `image t2i` [classic cohort, real 1376×768 JPEG] and `video t2v` veo-fast [`MEDIA_GENERATION_STATUS_SUCCESSFUL`, 11.5 MB `ftyp` mp4] both succeeded end-to-end at the new 1920×1080 viewport with every selector resolving [no drift]; SAPISIDHASH digest verified byte-identical [FIPS-safe `usedforsecurity=False`], auth unchanged). Prior: [v0.36.0](LIVE_VERIFICATION_v0.36.0.md) (**#316 diagnostic tooling live-verified against real Flow traffic** — `GFLOW_CLI_HAR_PATH` captured a real 54-entry HAR [real Flow domains, 34× 200s, and a genuine 401 the tool exists to diagnose]; `GFLOW_CLI_DEBUG_TRACEBACK`'s typed-error scoping confirmed live [console output identical with/without the flag on a `GFlowError`]; unhandled-path behavior covered by 6 new tests (30 total across both files) + SDD task review; #312 reference-entity-smuggling fix covered by its own dedicated e2e test). Prior: [v0.35.0](LIVE_VERIFICATION_v0.35.0.md) (**#305-follow-up multimodal reverse-engineering & storyboard tool proven live e2e** — `gflow tools run storyboard` successfully deconstructed a cyberpunk scene into a cohesive 4-panel sequence; `reverse-engineer` tool extracts frames from video streams/URLs using `claude-video`'s `watch.py` and produces structured prompt components using Gemini multimodal models; dynamic token budgeting and strict Pyright types verified via test suite). Prior: [v0.34.0](LIVE_VERIFICATION_v0.34.0.md) (**#299 `--ui-mode` gate proven live, ZERO credits** — drove `get_ui_driver` directly over 4 real editor loads, never generated: bidirectional switch works [classic→agentic via `_force_agent_mode`, agentic→classic via `_exit_agent_mode`, each DOM-reverified], `agentic`-when-already-agentic is idempotent, and a timed-out toggle → **exit-28 fail-fast before submission**; deprecation behavior-change [`prefer_classic`→exit 28] confirmed). Prior: [v0.33.0](LIVE_VERIFICATION_v0.33.0.md) (**#287-follow-up picker fix proven live on the original repro** — the v0.32.1-failing `i2v <uuid> --project <crowded>` command now exits 0 with a real 3.0 MB `ftyp` mp4, `frame_ref_attached` + `video_saved`, zero uploads; **#241 jitter live-verified on both paths** — flag `--jitter 4-6` → `batch_jitter_sleep seconds=4.98`, and the new 0.5–1.5 s default → `seconds=1.47`, each firing exactly once between two t2i submissions; environment note: the denon82 cohort is now natively agentic and round-2 generations hit the pre-existing #281 `MediaAttributionError`, not a regression). Prior: [v0.32.1](LIVE_VERIFICATION_v0.32.1.md) (**#293 teardown fix live-verified credit-free** — error-path aborts leave ZERO chrome.exe holding the profile [3× leak reproduced pre-fix same day]; force-close fallback + ProfileLockedError translation unit-pinned; #283 wave logic fixes mock-verified to exact call counts). Prior: [v0.32.0](LIVE_VERIFICATION_v0.32.0.md) (**#287 UUID i2v frames proven live e2e** — an in-project asset selected by media UUID in the frame-slot dialog [frame_ref_attached, startImage bound on the wire, ZERO upload events] produced a real 720×1280 interpolation, resolving the #237 negative prior; foreign UUID → exit 9 pre-generation with screenshot; **#288 duration fail-fast proven live** — exit 23 pre-submit saved a 10-credit generation, and the debug screenshot confirmed the duration control is ABSENT from the cohort's settings popover [locale hypothesis refuted]; upload-400 exit-27 repro deferred-with-reason — the original rejected JPEG no longer exists). Prior: [v0.31.0](LIVE_VERIFICATION_v0.31.0.md) (**#281 wrong-media defenses live-verified in BOTH directions, credit-free** — correct attribution downloads with real wire metadata [seed, 768×1376]; the incident-class lazy-render ambiguity reproduced live and hard-stopped with `MediaAttributionError` exit 26 naming every candidate UUID, no file written; **#282 two-UUID-ref i2i proven e2e** — both freshly generated portrait UUIDs picker-selected into one scene with both canonical identities visually transferred). Prior: [v0.30.0](LIVE_VERIFICATION_v0.30.0.md) (**agentic-cohort image path + both fixes live-verified credit-free** — native 768×1376 stills, i2i-chain + canonical-ref identity hold; the character-create `DataIntegrityError` and the `prefer_classic` WARNING were observed live then fixed [recorder idempotent + INFO log]; MCP video-param feature deferred-with-reason on Veo credits, per the stills-only goal). Prior: [v0.29.0](LIVE_VERIFICATION_v0.29.0.md) (**persistent `gflow instructions` CRUD proven live e2e GREEN, credit-free** — 11-command sequence on a fresh project: add incl. `--ref <image-UUID>` attach, list/enable/disable, master toggle both ways, TOML `apply` destructive full-sync with `enabled=false` honored, `rm` with server card-id preserved across mutations; movie brief-sync code-path-verified via the same primitives, deferred-with-reason on credits; MCP `gflow_instructions_*` tools are thin adapters over the live-verified primitives). Prior: [v0.28.0](LIVE_VERIFICATION_v0.28.0.md) (**agent instructions `-i` now steer agentic generation — crayon e2e GREEN**: a style-neutral prompt + an enabled "crayon" card produced an unmistakable crayon drawing through the real transport; root causes were conversational phrasing + the `project_brief.enabled` master switch). Prior: [v0.27.1](LIVE_VERIFICATION_v0.27.1.md) (**v0.27.0 release follow-up fixes and documentation sync**). Prior: [v0.27.0](LIVE_VERIFICATION_v0.27.0.md) (**`[style]` block with named variants + prompt-aware resume — credit-free CLI verification**: real `movie run --dry-run` 3-way ledger proving resolved-style display, `skip (done)` on matching `style_hash`, and `re-run (style changed)` after a style edit; composition is local, wire path untouched, so no Veo credits spent — first credited styled run will confirm visually). Prior: [v0.26.0](LIVE_VERIFICATION_v0.26.0.md) (**image i2i by generated-image UUID proven live e2e GREEN** — the existing asset is *selected in place* by its media UUID, no duplicate upload, producing an i2i output that references it; plus generated-image `display_name` capture, credited @C1ph3r404). Prior: [v0.25.0](LIVE_VERIFICATION_v0.25.0.md) (**#237 remote-UUID i2v proven live e2e GREEN** — a real 8s 720×1280 interpolation from a generated image's UUID, after the picker-search attach was reworked to local upload; #240 home-`.env` matrix live; plus two silent-failure guards root-caused live: video-as-image download rejection and rejected-upload fail-loud). Prior: [v0.24.0](LIVE_VERIFICATION_v0.24.0.md) (--project parity across CLI + MCP — automated coverage; live gen auth-gated). Prior: [v0.23.0](LIVE_VERIFICATION_v0.23.0.md) (MCP generation wiring proven live end-to-end — tool→worker→real Flow REST→structured-error translation; final image-write blocked by an expired session, environmental. #222 macOS fix reporter-verified e2e on Apple Silicon). Prior: [v0.22.0](LIVE_VERIFICATION_v0.22.0.md) (Tools framework — automated/CI coverage complete; credit/key-gated live feature run is a pending owner gate, see the doc) · [v0.21.0](LIVE_VERIFICATION_v0.21.0.md) (MCP server over stdio + HTTP/SSE — credit-free protocol handshake; 4 tools / 2 prompts / 3 resources confirmed on stdout) · [v0.20.1](LIVE_VERIFICATION_v0.20.1.md) · [v0.20.0](LIVE_VERIFICATION_v0.20.0.md) · [v0.19.0](LIVE_VERIFICATION_v0.19.0.md) · [v0.18.0](LIVE_VERIFICATION_v0.18.0.md) · [v0.17.0](LIVE_VERIFICATION_v0.17.0.md) · [v0.16.0](LIVE_VERIFICATION_v0.16.0.md) · [v0.13.0](LIVE_VERIFICATION_v0.13.0.md) · [v0.12.0](LIVE_VERIFICATION_v0.12.0.md) · [v0.11.0](LIVE_VERIFICATION_v0.11.0.md) · [v0.10.0](LIVE_VERIFICATION_v0.10.0.md) · [v0.9.1](LIVE_VERIFICATION_v0.9.1.md) · [v0.9.0](LIVE_VERIFICATION_v0.9.0.md) · [v0.8.1](LIVE_VERIFICATION_v0.8.1.md) · [v0.7.0](LIVE_VERIFICATION_v0.7.0.md)
 **"Where is the reverse-engineered wire protocol for a feature?"** → the `*_RECON.md` design docs: [CHARACTER_RECON](CHARACTER_RECON.md) (Flow character entity protocol), [IMAGE_UPSCALE_RECON](IMAGE_UPSCALE_RECON.md) (`/v1/flow/upsampleImage` wire). Naming convention: one `<FEATURE>_RECON.md` per reverse-engineered surface, kept as the durable spec after the feature ships. Pre-capture research recons: [ASSET_TAGGING_RECON](ASSET_TAGGING_RECON.md) (`@`-mention asset tagging — shipped in v0.40.0, see [LIVE_VERIFICATION_v0.40.0](LIVE_VERIFICATION_v0.40.0.md)).
 **"What was live-verified for the data layer (PR #58)?"** → [LIVE_VERIFICATION_data_layer](LIVE_VERIFICATION_data_layer.md) — 1 Imagen + 1 Veo credit on denon82, 6-layer ledger (file + magic + Pillow + DB rows + CLI round-trip + structlog)
 **"What was live-verified for the video-download feature (#29)?"** → [LIVE_VERIFICATION_video_download](LIVE_VERIFICATION_video_download.md)

{gflow_cli-0.41.0 → gflow_cli-0.42.0}/docs/LIVE_VERIFICATION_v0.40.0-production-readiness.md

@@ -15,6 +15,56 @@
 > and why, per the design spec's evidence rule
 > (`docs/superpowers/specs/2026-07-19-production-readiness-hardening-design.md`
 > § Evidence deliverable).
+>
+> **Update — see the "2026-07-21 live follow-up" section immediately below:** a
+> second live session closed the crash-recovery `remote_started` chain and the
+> D4 mid-launch cancellation path, added two durable live-gated e2e tests, and
+> corrected a false `flow_operation_id` invariant surfaced live. The lists in
+> this document describe the original v0.40.0 run; the follow-up section is the
+> authoritative update.
+
+## 2026-07-21 live follow-up (0.42.0 cycle)
+
+A second live session (profile `ffroliva`, 2026-07-21) closed several items left
+open by the original run and added durable live-gated e2e tests. `ffroliva` was
+on Flow's agentic UI cohort that day (classic mode unreachable, issue #174 A/B).
+
+**Crash-recovery `remote_started` chain — CONFIRMED (1 credit).** A real
+`veo-lite` t2v driven through the worker path with the C1 checkpoint observer
+completed, and the observer persisted a real `remote_started` checkpoint
+carrying the media handle to the DB — live-confirming the
+observer→checkpoint→DB chain for a *paid* generation. The recovery step
+(`remote_started` → `indeterminate`, never resubmit) is deterministic, proven by
+unit tests plus the credit-free phase-2 of the new e2e test. Durable artifact:
+`tests/e2e/test_crash_recovery_e2e.py` (live-gated, ~1 credit). Auto-reconciliation
+of a crashed paid job remains a deliberate **unimplemented** seam (the safe stub:
+`indeterminate` + manual reconcile).
+
+**Finding — `flow_operation_id` NULL on veo-lite → false invariant corrected.**
+That run also surfaced `operation_id=None` in the `remote_started` checkpoint for
+`veo-lite`, contradicting the prior claim that only omni-flash omits it.
+Root-caused: the operation-name parser is correct; `flow_operation_id` is
+best-effort/optional and `media_id` is the canonical handle (poll + download +
+every CLI lookup use it). The false invariant was corrected in `KNOWN_ISSUES.md`
+and the `client.py` docstring, and a regression test now locks the parser against
+the real veo fixture. No production behavior change — generation completed and
+downloaded correctly on the `None`.
+
+**D4 cancellation (mid-launch) — CONFIRMED (credit-free).** A real Chrome context
+was launched and the operation cancelled during `__aenter__` (before any submit
+gesture); the ProfileLease released (proven by successful re-acquisition) and no
+gflow-owned Chrome process remained. Durable artifact:
+`tests/e2e/test_cancellation_e2e.py` (live-gated, 0 credits). The other three D4
+cancellation sub-paths (mid-context-close with a wedged page, Ctrl-C during
+passive-capture reap, daemon SIGINT in-flight) remain deterministically
+unit-proven (`tests/api/test_concurrency.py`) and are not separately
+live-triggered — they are timing-fragile to reproduce live, and the mid-launch
+path exercises the same `run_teardown_step` release contract.
+
+**Still open (honest).** The full daemon/MCP lifecycle over HTTP (`gflow serve`)
+was not live-exercised (the worker path was exercised directly in the e2e test);
+auto-reconciliation remains unimplemented by design; the three fragile
+cancellation sub-paths remain unit-proven only.
 
 ## Commit and environment
 
@@ -265,12 +315,13 @@ step in the live sequence above ran to completion. What was deliberately
   run observed the handle capture but did not drive a crash-and-reconcile
   scenario live. Keep the underlying issue open; this is remaining work, not
   a regression.
-- **Remote macOS/Linux legs of the D2 `profile-lease-matrix` CI job** are
-  pending an authorized push — the POSIX `fcntl.flock` lease branch has not
-  been executed anywhere yet (Windows leg ran locally and is GREEN, see
-  Offline gates below). Per design-spec guidance, remote execution requires
-  explicit push/PR authorization and must not be reported as locally
-  executed.
+- **Remote macOS/Linux legs of the D2 `profile-lease-matrix` CI job**: this
+  gap is now CLOSED. PR #357 was pushed and its CI ran the
+  `profile-lease-matrix` job GREEN on all three OSes — `windows-latest`,
+  `macos-latest`, `ubuntu-latest` — exercising the POSIX `fcntl.flock` lease
+  branch on macOS/Linux for the first time (PR #357, `profile-lease-matrix`
+  job, merged 2026-07-20). This closes the CI-execution gap only; it is
+  CI-level evidence, not a live run against real Flow.
 - **omni-flash NULL-operation-id path** was not exercised — `veo-lite` was
   used deliberately for the T2V generation instead.
 - **D4's four live cancellation paths** (mid-launch cancel, mid-context.close
@@ -364,8 +415,12 @@ blocked paths" above for detail, not just offline-covered):
 - **Daemon/MCP live lifecycle and live queue-claim**: not exercised this
   pass — the live sequence used the direct `gflow image`/`gflow video`
   client path, not `gflow serve`/MCP.
-- **D2 remote CI legs (POSIX `fcntl` lease branch)**: pending an authorized
-  push; unexecuted anywhere in this branch's history.
+- **D2 remote CI legs (POSIX `fcntl` lease branch)**: CLOSED — PR #357 was
+  pushed and its CI ran the `profile-lease-matrix` job GREEN on all three
+  OSes (`windows-latest`, `macos-latest`, `ubuntu-latest`), exercising the
+  POSIX `fcntl.flock` lease branch on macOS/Linux (PR #357,
+  `profile-lease-matrix` job, merged 2026-07-20). CI-level evidence only —
+  not a live-Flow run.
 - **D4's four live cancellation paths**: unit-proven only, not driven live.
 - **omni-flash NULL-operation-id path**: not exercised (veo-lite used
   deliberately instead).

gflow_cli-0.42.0/docs/LIVE_VERIFICATION_v0.41.0.md

@@ -0,0 +1,57 @@
+# Live verification — v0.41.0 (2026-07-20)
+
+Release scope: **production-readiness hardening** ([#357]) — queue safety
+(versioned payloads, atomic claims, checkpointed execution phases),
+cross-process profile lease (`ProfileLockedError` exit 11),
+cancellation-safe browser teardown, driver honesty (typed
+`SupportsSendPrompt` injection, frozen `TransportSetup`), mention-index
+fail-closed (`MentionIndexUnavailableError` exit 29), external-CDP lifecycle
+removal, and nonfunctional `gflow video batch` removal.
+
+## Live matrix (ffroliva profile, 2026-07-20)
+
+Exercised against live Flow on `develop@a0c7a3f` (post-merge of #357):
+
+1. **Stale-session fail-fast**: authenticated session expired →
+   `AuthExpiredError` (exit 3) raised immediately, **no credits spent**.
+   Re-auth succeeded.
+2. **Free image generation**: `batchGenerateImages` returned HTTP 200, valid
+   768×1376 JPEG produced. Provenance recorded, zero leftover gflow-owned
+   Chrome processes, healthy cookie DB.
+3. **Paid veo-lite T2V**: `batchAsyncGenerateVideoText` returned HTTP 200,
+   `MEDIA_GENERATION_STATUS_SUCCESSFUL`, valid `ftyp` MP4 output. Clean
+   lease acquire/release/reacquire cycle observed.
+
+## 5-layer evidence ledger
+
+| Layer | Evidence |
+|---|---|
+| Stale-session | `AuthExpiredError` exit 3, no spend, re-auth succeeds |
+| Free image gen | `batchGenerateImages` 200, valid JPEG, provenance row |
+| Paid T2V | `batchAsyncGenerateVideoText` 200, `SUCCESSFUL`, valid MP4 |
+| Profile lease | Clean acquire/release/reacquire across runs, no `ProfileLockedError` |
+| Browser lifecycle | Zero leftover gflow-owned Chrome processes after each run |
+
+## Not verified live this cycle
+
+- Daemon/MCP live queue-claim and credit-free `remote_started`
+  re-entry reconciliation are offline-proven only.
+- The four D4 live cancellation paths are unit-proven.
+- omni-flash NULL-operation-id path not exercised (used veo-lite
+  deliberately).
+- The POSIX `fcntl.flock` lease branch has not executed anywhere yet — only
+  the Windows leg ran locally. A `profile-lease-matrix` CI job covers
+  Windows/macOS/Linux.
+
+## Pre-tag gates
+
+- Offline: repo hygiene, doc-links, `ruff check`, `ruff format --check`,
+  `pyright src` (0 errors) all clean.
+- Test suite: **2513 passed, 4 skipped**, 72 deselected.
+
+## Post-tag evidence
+
+- Tag: `v0.41.0`, signed (SSH), pushed to origin.
+- Release PR: [#358](https://github.com/ffroliva/gflow-cli/pull/358).
+
+[#357]: https://github.com/ffroliva/gflow-cli/pull/357

gflow_cli-0.42.0/docs/LIVE_VERIFICATION_v0.42.0.md

@@ -0,0 +1,30 @@
+# Live verification — v0.42.0
+
+Per the release protocol's live-verification gate, this records the live-verification
+status of v0.42.0's user-facing changes.
+
+## Content-safety 400 → ContentPolicyError (#359)
+
+**Status: not live-triggered — unit-covered.** Live-triggering this would require
+deliberately submitting a content-policy-violating prompt to Flow, which is
+inadvisable. The classification is covered by unit tests (`tests/test_errors_403.py`,
+`tests/test_errors_classification.py`), which assert that a content-safety `400`
+maps to `ContentPolicyError` (exit code 5) rather than `WireFormatError`.
+
+## Production-readiness live-gaps (#361)
+
+**Status: live-confirmed 2026-07-21** (profile `ffroliva`, 1 Veo credit). A real
+`veo-lite` t2v through the worker path persisted a real `remote_started` checkpoint
+(observer→checkpoint→DB chain confirmed), and crash-recovery classifies such a task
+`indeterminate` without resubmitting. Mid-launch cancellation released the ProfileLease
+and left no browser process behind (credit-free). Full evidence:
+[`docs/LIVE_VERIFICATION_v0.40.0-production-readiness.md`](LIVE_VERIFICATION_v0.40.0-production-readiness.md)
+§ "2026-07-21 live follow-up". Durable live-gated tests:
+`tests/e2e/test_crash_recovery_e2e.py`, `tests/e2e/test_cancellation_e2e.py`.
+
+## Documentation (#360, #362)
+
+**Status: n/a — no runtime feature.** Docs-only changes (Antigravity agent references;
+onboarding page; data anonymization; security contact → GitHub private vulnerability
+reporting). Verified by the documentation gates (`check_doc_links.py`,
+`test_documentation_gate.py`).

{gflow_cli-0.41.0 → gflow_cli-0.42.0}/docs/MCP.md

@@ -78,7 +78,7 @@ We analyzed whether a terminal-driven CLI guided by a text skill (e.g., `skills/
 | :--- | :--- | :--- |
 | **Output Fragility** | Ephemeral `stdout`/`stderr` log outputs are fragile to format adjustments, progress bars, and ANSI colors. | Strictly structured JSON payloads containing explicit metadata and absolute output file URIs. |
 | **Process Lifecycle**| High process startup overhead (Python import latency) on every individual execution. | Warm daemon process. Retains active caching of database metadata. |
-| **Concurrency** | Independent OS runs collide and crash the Chromium profile lock. | Serialized request execution via an internal `asyncio.Lock` queue. |
+| **Concurrency** | A second concurrent process against the same profile is rejected immediately by the cross-process `ProfileLease` with `ProfileLockedError` (exit code 11) — a clean fail-fast, never a crash, never a wait. Different profiles run fully in parallel. | Same `ProfileLease` enforcement applies to the daemon's requests — a same-profile call made while another is in flight (from the CLI or the MCP daemon) is rejected immediately with `ProfileLockedError` rather than queued or crashed; different profiles still run fully in parallel. |
 | **Error Handling** | Agent must scan logs for strings or parse exit codes to check status. | Strongly-typed JSON-RPC errors with mapped codes and clear remediation. |
 | **Transport Safety** | Volatile console printing. | Stdout is strictly isolated for JSON-RPC; all logs and warnings route to stderr. |
 

{gflow_cli-0.41.0 → gflow_cli-0.42.0}/docs/PROJECT_STATUS.md

@@ -4,7 +4,7 @@
 
 ## Current release
 
-**v0.41.0 — alpha.** **Production-readiness hardening** ([#357]): queue safety (versioned payloads, atomic claims, checkpointed execution phases), cross-process profile lease (`ProfileLockedError` exit 11), cancellation-safe browser teardown, driver honesty (typed `SupportsSendPrompt` injection, frozen `TransportSetup`), mention-index fail-closed (`MentionIndexUnavailableError` exit 29), and external-CDP lifecycle removal. Removed nonfunctional `gflow video batch` command. Also: `/gflow:live-verify` skill for per-feature live-verification enforcement. 2513 tests pass; live-verified against real Flow (stale-session fail-fast, free image gen, paid veo-lite T2V). See [LIVE_VERIFICATION_v0.40.0-production-readiness.md](LIVE_VERIFICATION_v0.40.0-production-readiness.md).
+**v0.41.0 — alpha.** **Production-readiness hardening** ([#357]): queue safety (versioned payloads, atomic claims, checkpointed execution phases), cross-process profile lease (`ProfileLockedError` exit 11), cancellation-safe browser teardown, driver honesty (typed `SupportsSendPrompt` injection, frozen `TransportSetup`), mention-index fail-closed (`MentionIndexUnavailableError` exit 29), and external-CDP lifecycle removal. Removed nonfunctional manifest-driven video batch command (never worked end-to-end; loop `gflow video t2v`/`i2v` from the shell instead). Also: `/gflow:live-verify` skill for per-feature live-verification enforcement. 2513 tests pass; live-verified against real Flow (stale-session fail-fast, free image gen, paid veo-lite T2V). See [LIVE_VERIFICATION_v0.40.0-production-readiness.md](LIVE_VERIFICATION_v0.40.0-production-readiness.md).
 
 **Develop (unreleased, post-v0.41.0):** *(empty — develop is the staging branch for the next release).*
 
@@ -16,7 +16,7 @@
 
 <details><summary>v0.39.0 — failed-generation persistence + gflow data list errors</summary>
 
-**v0.39.0 — alpha.** **Failed generations are now persisted to the local catalog (#341).** Every paid-generation path (`video t2v/i2v/r2v`, `video chain`, `image t2i/i2i`, multi-prompt t2i, `gflow run`, `image batch`, `movie run`, and the async worker) now records a terminal `status="failed"` operation row — with a stable `error_type` derived from the exception's RFC 9457 `problem_type` and a redacted `error_detail` — before the error propagates, so WAF-403 block onset, duration, and recovery windows are measurable instead of reconstructed from memory. The new `gflow data list errors` subcommand browses failed operations newest-first, and videos whose poll returns `succeeded=false` are now recorded as `failed` (previously mis-recorded as `succeeded`). Bounded retention + export deferred to #345. Verification: [LIVE_VERIFICATION_v0.39.0](LIVE_VERIFICATION_v0.39.0.md) (a real `image t2i` HTTP-400 wire failure produced the first `failed` row in the production catalog, $0).
+**v0.39.0 — alpha.** **Failed generations are now persisted to the local catalog (#341).** Every paid-generation path (`video t2v/i2v/r2v`, `video chain`, `image t2i/i2i`, multi-prompt t2i, `gflow run`, `movie run`, and the async worker) now records a terminal `status="failed"` operation row — with a stable `error_type` derived from the exception's RFC 9457 `problem_type` and a redacted `error_detail` — before the error propagates, so WAF-403 block onset, duration, and recovery windows are measurable instead of reconstructed from memory. The new `gflow data list errors` subcommand browses failed operations newest-first, and videos whose poll returns `succeeded=false` are now recorded as `failed` (previously mis-recorded as `succeeded`). Bounded retention + export deferred to #345. Verification: [LIVE_VERIFICATION_v0.39.0](LIVE_VERIFICATION_v0.39.0.md) (a real `image t2i` HTTP-400 wire failure produced the first `failed` row in the production catalog, $0).
 
 </details>
 
@@ -250,7 +250,7 @@ reporter-verified e2e on macOS).
 | Agentic-pin recovery: opt-in reload after a real Agent-toggle click (#338) | ✅ done (v0.38.1) |
 | Failed generations persisted to the local catalog + `gflow data list errors` (#341) | ✅ done (v0.39.0) |
 | Prompt `@`-mention resolution for asset tagging (#344) | ✅ done (v0.40.0) |
-| Manifest-driven video batch runner on `ui_automation` | ❌ removed — never worked end-to-end, shipped as a nonfunctional stub; see [CHANGELOG § Unreleased](../CHANGELOG.md#unreleased). For multi-clip video, loop `gflow video t2v`/`i2v` from the shell; `gflow image batch` remains supported for images. |
+| Manifest-driven video batch runner on `ui_automation` | ❌ removed — never worked end-to-end, shipped as a nonfunctional stub; see v0.41.0 changelog. For multi-clip video, loop `gflow video t2v`/`i2v` from the shell; `gflow image batch` remains supported for images. |
 | Persistence layer (stay-mounted batch sessions across project boundaries) | ⏳ Phase B |
 | Provider abstraction for official Veo 3.1 API | ⏳ planned |
 | Signed-tag CI verification automation (no manual signing in CI yet) | ⏳ planned |

{gflow_cli-0.41.0 → gflow_cli-0.42.0}/docs/superpowers/specs/2026-07-19-live-verify-design.md

@@ -76,7 +76,7 @@ gflow-cli work done, especially anything touching a generation path — Part 2")
 proactive-invocation mechanism fires it at both points, the same way `brainstorming` already
 auto-triggers on "let's build X." This mechanism is specific to agents that honor the
 `using-superpowers` convention (this project's primary Claude Code workflow) — it is not a
-substitute for discoverability by other agents this repo explicitly supports (Cursor/Codex/Gemini
+substitute for discoverability by other agents this repo explicitly supports (Cursor/Codex/Antigravity
 CLI/Aider), which is why the `AGENTS.md` skills-reference table row (3.5) is required, not optional
 reinforcement.
 

{gflow_cli-0.41.0 → gflow_cli-0.42.0}/llms.txt

@@ -24,7 +24,7 @@ This file is forward-staged for a future docs site at https://ffroliva.github.io
 
 ## For AI coding agents
 
-- [AGENTS.md](https://github.com/ffroliva/gflow-cli/blob/main/AGENTS.md): Universal coding-agent spec (Cursor, Codex, Aider, Gemini CLI, Claude Code, etc.).
+- [AGENTS.md](https://github.com/ffroliva/gflow-cli/blob/main/AGENTS.md): Universal coding-agent spec (Cursor, Codex, Aider, Antigravity, Claude Code, etc.).
 - [CLAUDE.md](https://github.com/ffroliva/gflow-cli/blob/main/CLAUDE.md): Claude Code's auto-loaded memory hub.
 - [docs/AGENT_GUIDE.md](https://github.com/ffroliva/gflow-cli/blob/main/docs/AGENT_GUIDE.md): Mandates and routing rules for agents.
 

{gflow_cli-0.41.0 → gflow_cli-0.42.0}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "gflow-cli"
-version = "0.41.0"
+version = "0.42.0"
 description = "Unofficial CLI for Google Flow — drive Veo image-to-video generations from the terminal."
 readme = "README.md"
 license = { file = "LICENSE" }

{gflow_cli-0.41.0 → gflow_cli-0.42.0}/scripts/ci/check_materiality.py

@@ -95,7 +95,7 @@ def build_report(material: list[str], routine: list[str], signals: dict[str, boo
             "### Recommended gates",
             "- Run **`/gflow:predict`** (Claude Code), **or**",
             "- Read **`skills/predict/SKILL.md`** and produce the 5-persona "
-            "verdict yourself (Cursor / Codex / Gemini / Aider), **or**",
+            "verdict yourself (Cursor / Codex / Antigravity / Aider), **or**",
             "- A human reviewer documents the risk assessment in the PR.",
             "",
             "Then a council pass: `/gflow:pr-council-review` or read "