feat: Vendor hosted Stage into the CLI: generation parity, review UI, and configurability

[dastratakos]

Summary

The hosted product (ReviewStage/stage) is sunsetted; this PR vendors its unported functionality into the CLI so users can continue on stagereview. Everything is ported as closely as possible to the hosted implementation (code, prompts, constants, and tests copied verbatim where the architectures align), based on a six-dimension gap analysis of the two repos' code and git histories.

Changes

Chapter generation — prompts (skills/stage-chapters/SKILL.md)

• Full risk-classification rubric with riskLevel/riskReasons outputs (hosted #896)
• Prologue rootCause section (hosted #1051), chapter-level mermaid diagram guidance (hosted #771)
• Restored hosted's full exemplar density, PR title/body + additional-instructions context, keyChange inline-markdown rule

Chapter generation — pipeline & data model (packages/cli, packages/types)

• No-newline markers filtered from parsed hunks (hosted #1132) with regression tests
• Hosted ignore semantics: full DEFAULT_IGNORE_PATTERNS (7 new lockfiles, #871), one case-insensitive matcher so .stageignore! negations re-include defaults
• Zero-hunk files (pure renames) included in Other Changes via the header-only sentinel
• riskLevel/riskReasons + rootCause across types, DB (+migration), ingestion, routes
• What-to-Review ordering by file order then start line (#865/#970); dense 0-based chapterIndex
• prep: PULL REQUEST (title/body, 4k cap), STATS, and ADDITIONAL INSTRUCTIONS sections (.stageinstructions file + --instructions flag, hosted #1036 analog)
• Repair-not-reject line refs; pre-assembled chapter files no longer bypass validation; literal-\n unescaping; consistent diff prefixes

Review UI (packages/web)

• Activity tab: conversation timeline (gh-api-backed), review cards, review-comment diff previews (#1105), events column, read-only reactions
• Continuous chapter review mode (Page/Scroll, #1083) with active-chapter preservation on mode switch
• Risk badges, per-chapter/per-file comment counts, remaining-review-lines sticky header (#1107)
• Markdown stack: shiki highlighting (#796/#826), mermaid in chapter summaries (#771), GitHub image proxy for private attachments (#820)
• Diff extras: image diffs (base64 over the wire), full previews for moved files (#1106), virtualized file list, minimize inline comments
• Small fixes: bot reviewer badges (#1131), resolve-button placement (#1133), sticky-header z-index (#1019), CI popover scroll (#1061), min-height floor (#1014), annotation tinting (#757/#1016), index j/k (#835), dark default theme (#819), derived chapter viewed state (#1021)
• GitHub: stale pending-review recovery (#1072), label management (#1071)
• GitHub viewed-file sync: marks/unmarks files on GitHub per hosted's rules and unions GitHub's viewed state into the local view state (PR runs, best-effort)
• Merge: ruleset-aware allowed merge methods (#1057), single inline merge action with status chip and queue position (#1052)
• Stacked-PR detection and navigation with ⌘⇧↑/⌘⇧↓ (#1049); entries with a local run navigate to it, others link to GitHub
• Review panel keyboard shortcuts: ⌘J toggle, ⌘↩ submit

Configurability

• Diff typography (font/size/line height/ligatures, #974), per-app-theme syntax themes (62), app text size (#1038), chapter panel position (#898), What-to-Review visibility (#1077)

Descoped (depend on hosted backends): Stagent chat, ⌘K action board, reaction toggling, timeline-side comment mutations, mobile layouts, AI chapter editing/regeneration UI.

Testing

• pnpm typecheck, pnpm lint, pnpm test all green (73 files, 643 tests — including regression suites ported from hosted for the diff parser, file filtering, other-changes, line-ref repair, remaining-line counts, timeline assembly, full-file previews, and review recovery)
• pnpm build (SPA + CLI bundle) succeeds

---

Open workspace in Conductor

---

Summary by cubic

Vendors the hosted Stage experience into the CLI so teams can keep using stagereview after the hosted app sunset, with generation parity, a full review UI, GitHub viewed-file sync, labels/stack/merge parity, and configurable display/settings. Also aligns the continuous chapter panel header with paged mode by reusing the shared chapter navigator, with in-page navigation to avoid remounts.

• Bug Fixes

  • Drop stale comment‑draft store writes that settle after stack navigation.
  • Count outdated GitHub review threads via their original coordinates and carry them as a null‑anchored variant so file and chapter comment badges stay accurate.
  • Restore j/k navigation and scroll‑into‑view by typing the --content-top CSS variable, fixing keyboard and deep‑link scrolling in Files and chapter pages.

• Refactors

  • Drop the legacy syntax‑theme localStorage migration; the curated single‑key setting now resets to defaults on upgrade. Remove hosted‑Stage provenance comments (no behavior change).

^{Written for commit 0f37e90. Summary will update on new commits.}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Bugbot skipped with an unresolved finding, Security Agent reported an unresolved high-severity issue, and this PR is above the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit:5d995b7d33

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[cubic-dev-ai]

All reported issues were addressed across 128 files

<sub>Reply with feedback, questions, or to request a fix.

Re-trigger cubic</sub>

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Cursor Bugbot skipped, an unresolved high-severity Security Agent finding remains, unresolved review threads are still open, and this PR exceeds the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cubic-dev-ai]

All reported issues were addressed across 21 files (changes from recent commits).

<sub>Reply with feedback, questions, or to request a fix.

Re-trigger cubic</sub>

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit:570505304f

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Bugbot completed as skipped with an unresolved finding, an unresolved high-severity Security Agent finding remains, and unresolved review threads remain above the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Bugbot skipped with an unresolved finding, an unresolved high-severity Security Agent finding remains, unresolved review threads remain, and this PR exceeds the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cursor]

Cursor Bugbot has reviewed your changes using default effort and found 1 potential issue.

There are 2 total unresolved issues (including 1 from previous review).

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit ba5b0d2. Configure here.}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped with an unresolved high-severity finding still open, and this PR exceeds the medium approval threshold (migrations, GitHub/token-backed routes, lockfile, cross-cutting CLI/web). Human review is needed; assigning reviewers.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit:ba5b0d2b1e

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[cursor]

Cursor Bugbot has reviewed your changes using default effort and found 1 potential issue.

There are 2 total unresolved issues (including 1 from previous review).

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 7b9b3e7. Configure here.}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped with an unresolved finding still open, and this PR exceeds the medium approval threshold (migrations, GitHub/token-backed routes, cross-cutting CLI/web). Human review is needed; assigning reviewers.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cubic-dev-ai]

1 issue found across 5 files (changes from recent commits).

Prompt for AI agents (unresolved issues)

Check if these issues are valid — if so, understand the root cause of each and fix them. If appropriate, use sub-agents to investigate and fix each issue separately.
<file name="packages/cli/src/__tests__/view-state-github-harness.ts">
<violation number="1" location="packages/cli/src/__tests__/view-state-github-harness.ts:130">
P3: The shim hardcodes state OPEN on every pr list response, so the routes tests using this harness cannot exercise the closed/merged-PR fallback in getPullRequestOrThrow (pull-request.ts:196, listParsed.data.find(pr => pr.state === OPEN) ?? listParsed.data[0]). The available options (branchPrNumber or null) only cover the open-PR and no-PR cases, leaving the newest-closed-PR fallback and prefer-open ordering untested by the integration suite. Consider exposing a configurable pr-list entries option on the harness so that fallback gets regression coverage.</violation>
</file>

<sub>Tip: Review your code locally with the cubic CLI to iterate faster.

Re-trigger cubic</sub>

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped, and this PR exceeds the medium approval threshold (migrations, GitHub/token-backed routes, lockfile, cross-cutting CLI/web). Human review is needed; assigning reviewers.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped, and this PR exceeds the medium approval threshold (migrations, GitHub/token-backed routes, lockfile, cross-cutting CLI/web). Human review is needed; assigning reviewers.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cubic-dev-ai]

All reported issues were addressed across 8 files (changes from recent commits).

<sub>Tip: Review your code locally with the cubic CLI to iterate faster.

Re-trigger cubic</sub>

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit:cd06993e8c

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Cursor Bugbot completed as skipped, and this PR exceeds the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cubic-dev-ai]

All reported issues were addressed across 8 files (changes from recent commits).

<sub>Tip: Review your code locally with the cubic CLI to iterate faster.

Re-trigger cubic</sub>

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Cursor Bugbot completed as skipped, and this PR exceeds the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cursor]

Cursor Bugbot has reviewed your changes using default effort and found 1 potential issue.

There are 2 total unresolved issues (including 1 from previous review).

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit acfe195. Configure here.}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped with an unresolved finding still open, and this PR exceeds the medium approval threshold. Human review is needed; assigning reviewers.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cubic-dev-ai]

All reported issues were addressed across 72 files (changes from recent commits).

<sub>Tip: Review your code locally with the cubic CLI to iterate faster.

Re-trigger cubic</sub>

[cursor]

Cursor Bugbot has reviewed your changes using default effort and found 1 potential issue.

There are 2 total unresolved issues (including 1 from previous review).

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 0f37e90. Configure here.}

[cursor]

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped with unresolved findings still open, and this PR exceeds the medium approval threshold. Human review is needed; a reviewer is already assigned.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[dastratakos]

Summary

The hosted product (ReviewStage/stage) is sunsetted; this PR vendors its unported functionality into the CLI so users can continue on stagereview. Everything is ported as closely as possible to the hosted implementation (code, prompts, constants, and tests copied verbatim where the architectures align), based on a six-dimension gap analysis of the two repos' code and git histories.

Changes

Chapter generation — prompts (skills/stage-chapters/SKILL.md)

• Full risk-classification rubric with riskLevel/riskReasons outputs (hosted #896)
• Prologue rootCause section (hosted #1051), chapter-level mermaid diagram guidance (hosted #771)
• Restored hosted's full exemplar density, PR title/body + additional-instructions context, keyChange inline-markdown rule

Chapter generation — pipeline & data model (packages/cli, packages/types)

• No-newline markers filtered from parsed hunks (hosted #1132) with regression tests
• Hosted ignore semantics: full DEFAULT_IGNORE_PATTERNS (7 new lockfiles, #871), one case-insensitive matcher so .stageignore! negations re-include defaults
• Zero-hunk files (pure renames) included in Other Changes via the header-only sentinel
• riskLevel/riskReasons + rootCause across types, DB (+migration), ingestion, routes
• What-to-Review ordering by file order then start line (#865/#970); dense 0-based chapterIndex
• prep: PULL REQUEST (title/body, 4k cap), STATS, and ADDITIONAL INSTRUCTIONS sections (.stageinstructions file + --instructions flag, hosted #1036 analog)
• Repair-not-reject line refs; pre-assembled chapter files no longer bypass validation; literal-\n unescaping; consistent diff prefixes

Review UI (packages/web)

• Activity tab: conversation timeline (gh-api-backed), review cards, review-comment diff previews (#1105), events column, read-only reactions
• Continuous chapter review mode (Page/Scroll, #1083) with active-chapter preservation on mode switch
• Risk badges, per-chapter/per-file comment counts, remaining-review-lines sticky header (#1107)
• Markdown stack: shiki highlighting (#796/#826), mermaid in chapter summaries (#771), GitHub image proxy for private attachments (#820)
• Diff extras: image diffs (base64 over the wire), full previews for moved files (#1106), virtualized file list, minimize inline comments
• Small fixes: bot reviewer badges (#1131), resolve-button placement (#1133), sticky-header z-index (#1019), CI popover scroll (#1061), min-height floor (#1014), annotation tinting (#757/#1016), index j/k (#835), dark default theme (#819), derived chapter viewed state (#1021)
• GitHub: stale pending-review recovery (#1072), label management (#1071)
• GitHub viewed-file sync: marks/unmarks files on GitHub per hosted's rules and unions GitHub's viewed state into the local view state (PR runs, best-effort)
• Merge: ruleset-aware allowed merge methods (#1057), single inline merge action with status chip and queue position (#1052)
• Stacked-PR detection and navigation with ⌘⇧↑/⌘⇧↓ (#1049); entries with a local run navigate to it, others link to GitHub
• Review panel keyboard shortcuts: ⌘J toggle, ⌘↩ submit

Configurability

• Diff typography (font/size/line height/ligatures, #974), per-app-theme syntax themes (62), app text size (#1038), chapter panel position (#898), What-to-Review visibility (#1077)

Descoped (depend on hosted backends): Stagent chat, ⌘K action board, reaction toggling, timeline-side comment mutations, mobile layouts, AI chapter editing/regeneration UI.

Testing

• pnpm typecheck, pnpm lint, pnpm test all green (73 files, 643 tests — including regression suites ported from hosted for the diff parser, file filtering, other-changes, line-ref repair, remaining-line counts, timeline assembly, full-file previews, and review recovery)
• pnpm build (SPA + CLI bundle) succeeds

---

Open workspace in Conductor

---

Summary by cubic

Vendors the hosted Stage experience into the CLI so teams can keep using stagereview after the hosted app sunset, with generation parity, a full review UI, GitHub viewed-file sync, labels/stack/merge parity, and configurable display/settings. Also aligns the continuous chapter panel header with paged mode by reusing the shared chapter navigator, with in-page navigation to avoid remounts.

• Bug Fixes

  • Drop stale comment‑draft store writes that settle after stack navigation.
  • Count outdated GitHub review threads via their original coordinates and carry them as a null‑anchored variant so file and chapter comment badges stay accurate.
  • Restore j/k navigation and scroll‑into‑view by typing the --content-top CSS variable, fixing keyboard and deep‑link scrolling in Files and chapter pages.

• Refactors

  • Drop the legacy syntax‑theme localStorage migration; the curated single‑key setting now resets to defaults on upgrade. Remove hosted‑Stage provenance comments (no behavior change).

^{Written for commit 0f37e90. Summary will update on new commits.}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Bugbot skipped with an unresolved finding, Security Agent reported an unresolved high-severity issue, and this PR is above the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit:5d995b7d33

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[cubic-dev-ai]

All reported issues were addressed across 128 files

<sub>Reply with feedback, questions, or to request a fix.

Re-trigger cubic</sub>

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Cursor Bugbot skipped, an unresolved high-severity Security Agent finding remains, unresolved review threads are still open, and this PR exceeds the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cubic-dev-ai]

All reported issues were addressed across 21 files (changes from recent commits).

<sub>Reply with feedback, questions, or to request a fix.

Re-trigger cubic</sub>

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit:570505304f

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Bugbot completed as skipped with an unresolved finding, an unresolved high-severity Security Agent finding remains, and unresolved review threads remain above the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Bugbot skipped with an unresolved finding, an unresolved high-severity Security Agent finding remains, unresolved review threads remain, and this PR exceeds the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cursor]

Cursor Bugbot has reviewed your changes using default effort and found 1 potential issue.

There are 2 total unresolved issues (including 1 from previous review).

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit ba5b0d2. Configure here.}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped with an unresolved high-severity finding still open, and this PR exceeds the medium approval threshold (migrations, GitHub/token-backed routes, lockfile, cross-cutting CLI/web). Human review is needed; assigning reviewers.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit:ba5b0d2b1e

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[cursor]

Cursor Bugbot has reviewed your changes using default effort and found 1 potential issue.

There are 2 total unresolved issues (including 1 from previous review).

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 7b9b3e7. Configure here.}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped with an unresolved finding still open, and this PR exceeds the medium approval threshold (migrations, GitHub/token-backed routes, cross-cutting CLI/web). Human review is needed; assigning reviewers.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cubic-dev-ai]

1 issue found across 5 files (changes from recent commits).

Prompt for AI agents (unresolved issues)

Check if these issues are valid — if so, understand the root cause of each and fix them. If appropriate, use sub-agents to investigate and fix each issue separately.
<file name="packages/cli/src/__tests__/view-state-github-harness.ts">
<violation number="1" location="packages/cli/src/__tests__/view-state-github-harness.ts:130">
P3: The shim hardcodes state OPEN on every pr list response, so the routes tests using this harness cannot exercise the closed/merged-PR fallback in getPullRequestOrThrow (pull-request.ts:196, listParsed.data.find(pr => pr.state === OPEN) ?? listParsed.data[0]). The available options (branchPrNumber or null) only cover the open-PR and no-PR cases, leaving the newest-closed-PR fallback and prefer-open ordering untested by the integration suite. Consider exposing a configurable pr-list entries option on the harness so that fallback gets regression coverage.</violation>
</file>

<sub>Tip: Review your code locally with the cubic CLI to iterate faster.

Re-trigger cubic</sub>

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped, and this PR exceeds the medium approval threshold (migrations, GitHub/token-backed routes, lockfile, cross-cutting CLI/web). Human review is needed; assigning reviewers.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped, and this PR exceeds the medium approval threshold (migrations, GitHub/token-backed routes, lockfile, cross-cutting CLI/web). Human review is needed; assigning reviewers.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cubic-dev-ai]

All reported issues were addressed across 8 files (changes from recent commits).

<sub>Tip: Review your code locally with the cubic CLI to iterate faster.

Re-trigger cubic</sub>

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit:cd06993e8c

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Cursor Bugbot completed as skipped, and this PR exceeds the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cubic-dev-ai]

All reported issues were addressed across 8 files (changes from recent commits).

<sub>Tip: Review your code locally with the cubic CLI to iterate faster.

Re-trigger cubic</sub>

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Cursor Bugbot completed as skipped, and this PR exceeds the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cursor]

Cursor Bugbot has reviewed your changes using default effort and found 1 potential issue.

There are 2 total unresolved issues (including 1 from previous review).

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit acfe195. Configure here.}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped with an unresolved finding still open, and this PR exceeds the medium approval threshold. Human review is needed; assigning reviewers.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cubic-dev-ai]

All reported issues were addressed across 72 files (changes from recent commits).

<sub>Tip: Review your code locally with the cubic CLI to iterate faster.

Re-trigger cubic</sub>

[cursor]

Cursor Bugbot has reviewed your changes using default effort and found 1 potential issue.

There are 2 total unresolved issues (including 1 from previous review).

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 0f37e90. Configure here.}

[cursor]

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped with unresolved findings still open, and this PR exceeds the medium approval threshold. Human review is needed; a reviewer is already assigned.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[dastratakos]

Summary

The hosted product (ReviewStage/stage) is sunsetted; this PR vendors its unported functionality into the CLI so users can continue on stagereview. Everything is ported as closely as possible to the hosted implementation (code, prompts, constants, and tests copied verbatim where the architectures align), based on a six-dimension gap analysis of the two repos' code and git histories.

Changes

Chapter generation — prompts (skills/stage-chapters/SKILL.md)

• Full risk-classification rubric with riskLevel/riskReasons outputs (hosted #896)
• Prologue rootCause section (hosted #1051), chapter-level mermaid diagram guidance (hosted #771)
• Restored hosted's full exemplar density, PR title/body + additional-instructions context, keyChange inline-markdown rule

Chapter generation — pipeline & data model (packages/cli, packages/types)

• No-newline markers filtered from parsed hunks (hosted #1132) with regression tests
• Hosted ignore semantics: full DEFAULT_IGNORE_PATTERNS (7 new lockfiles, #871), one case-insensitive matcher so .stageignore! negations re-include defaults
• Zero-hunk files (pure renames) included in Other Changes via the header-only sentinel
• riskLevel/riskReasons + rootCause across types, DB (+migration), ingestion, routes
• What-to-Review ordering by file order then start line (#865/#970); dense 0-based chapterIndex
• prep: PULL REQUEST (title/body, 4k cap), STATS, and ADDITIONAL INSTRUCTIONS sections (.stageinstructions file + --instructions flag, hosted #1036 analog)
• Repair-not-reject line refs; pre-assembled chapter files no longer bypass validation; literal-\n unescaping; consistent diff prefixes

Review UI (packages/web)

• Activity tab: conversation timeline (gh-api-backed), review cards, review-comment diff previews (#1105), events column, read-only reactions
• Continuous chapter review mode (Page/Scroll, #1083) with active-chapter preservation on mode switch
• Risk badges, per-chapter/per-file comment counts, remaining-review-lines sticky header (#1107)
• Markdown stack: shiki highlighting (#796/#826), mermaid in chapter summaries (#771), GitHub image proxy for private attachments (#820)
• Diff extras: image diffs (base64 over the wire), full previews for moved files (#1106), virtualized file list, minimize inline comments
• Small fixes: bot reviewer badges (#1131), resolve-button placement (#1133), sticky-header z-index (#1019), CI popover scroll (#1061), min-height floor (#1014), annotation tinting (#757/#1016), index j/k (#835), dark default theme (#819), derived chapter viewed state (#1021)
• GitHub: stale pending-review recovery (#1072), label management (#1071)
• GitHub viewed-file sync: marks/unmarks files on GitHub per hosted's rules and unions GitHub's viewed state into the local view state (PR runs, best-effort)
• Merge: ruleset-aware allowed merge methods (#1057), single inline merge action with status chip and queue position (#1052)
• Stacked-PR detection and navigation with ⌘⇧↑/⌘⇧↓ (#1049); entries with a local run navigate to it, others link to GitHub
• Review panel keyboard shortcuts: ⌘J toggle, ⌘↩ submit

Configurability

• Diff typography (font/size/line height/ligatures, #974), per-app-theme syntax themes (62), app text size (#1038), chapter panel position (#898), What-to-Review visibility (#1077)

Descoped (depend on hosted backends): Stagent chat, ⌘K action board, reaction toggling, timeline-side comment mutations, mobile layouts, AI chapter editing/regeneration UI.

Testing

• pnpm typecheck, pnpm lint, pnpm test all green (73 files, 643 tests — including regression suites ported from hosted for the diff parser, file filtering, other-changes, line-ref repair, remaining-line counts, timeline assembly, full-file previews, and review recovery)
• pnpm build (SPA + CLI bundle) succeeds

---

Open workspace in Conductor

---

Summary by cubic

Vendors the hosted Stage experience into the CLI so teams can keep using stagereview after the hosted app sunset, with generation parity, a full review UI, GitHub viewed-file sync, labels/stack/merge parity, and configurable display/settings. Also aligns the continuous chapter panel header with paged mode by reusing the shared chapter navigator, with in-page navigation to avoid remounts.

• Bug Fixes

  • Drop stale comment‑draft store writes that settle after stack navigation.
  • Count outdated GitHub review threads via their original coordinates and carry them as a null‑anchored variant so file and chapter comment badges stay accurate.
  • Restore j/k navigation and scroll‑into‑view by typing the --content-top CSS variable, fixing keyboard and deep‑link scrolling in Files and chapter pages.

• Refactors

  • Drop the legacy syntax‑theme localStorage migration; the curated single‑key setting now resets to defaults on upgrade. Remove hosted‑Stage provenance comments (no behavior change).

^{Written for commit 0f37e90. Summary will update on new commits.}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Bugbot skipped with an unresolved finding, Security Agent reported an unresolved high-severity issue, and this PR is above the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit:5d995b7d33

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[cubic-dev-ai]

All reported issues were addressed across 128 files

<sub>Reply with feedback, questions, or to request a fix.

Re-trigger cubic</sub>

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Cursor Bugbot skipped, an unresolved high-severity Security Agent finding remains, unresolved review threads are still open, and this PR exceeds the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cubic-dev-ai]

All reported issues were addressed across 21 files (changes from recent commits).

<sub>Reply with feedback, questions, or to request a fix.

Re-trigger cubic</sub>

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit:570505304f

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Bugbot completed as skipped with an unresolved finding, an unresolved high-severity Security Agent finding remains, and unresolved review threads remain above the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Bugbot skipped with an unresolved finding, an unresolved high-severity Security Agent finding remains, unresolved review threads remain, and this PR exceeds the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cursor]

Cursor Bugbot has reviewed your changes using default effort and found 1 potential issue.

There are 2 total unresolved issues (including 1 from previous review).

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit ba5b0d2. Configure here.}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped with an unresolved high-severity finding still open, and this PR exceeds the medium approval threshold (migrations, GitHub/token-backed routes, lockfile, cross-cutting CLI/web). Human review is needed; assigning reviewers.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit:ba5b0d2b1e

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[cursor]

Cursor Bugbot has reviewed your changes using default effort and found 1 potential issue.

There are 2 total unresolved issues (including 1 from previous review).

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 7b9b3e7. Configure here.}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped with an unresolved finding still open, and this PR exceeds the medium approval threshold (migrations, GitHub/token-backed routes, cross-cutting CLI/web). Human review is needed; assigning reviewers.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cubic-dev-ai]

1 issue found across 5 files (changes from recent commits).

Prompt for AI agents (unresolved issues)

Check if these issues are valid — if so, understand the root cause of each and fix them. If appropriate, use sub-agents to investigate and fix each issue separately.
<file name="packages/cli/src/__tests__/view-state-github-harness.ts">
<violation number="1" location="packages/cli/src/__tests__/view-state-github-harness.ts:130">
P3: The shim hardcodes state OPEN on every pr list response, so the routes tests using this harness cannot exercise the closed/merged-PR fallback in getPullRequestOrThrow (pull-request.ts:196, listParsed.data.find(pr => pr.state === OPEN) ?? listParsed.data[0]). The available options (branchPrNumber or null) only cover the open-PR and no-PR cases, leaving the newest-closed-PR fallback and prefer-open ordering untested by the integration suite. Consider exposing a configurable pr-list entries option on the harness so that fallback gets regression coverage.</violation>
</file>

<sub>Tip: Review your code locally with the cubic CLI to iterate faster.

Re-trigger cubic</sub>

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped, and this PR exceeds the medium approval threshold (migrations, GitHub/token-backed routes, lockfile, cross-cutting CLI/web). Human review is needed; assigning reviewers.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped, and this PR exceeds the medium approval threshold (migrations, GitHub/token-backed routes, lockfile, cross-cutting CLI/web). Human review is needed; assigning reviewers.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cubic-dev-ai]

All reported issues were addressed across 8 files (changes from recent commits).

<sub>Tip: Review your code locally with the cubic CLI to iterate faster.

Re-trigger cubic</sub>

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit:cd06993e8c

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Cursor Bugbot completed as skipped, and this PR exceeds the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cubic-dev-ai]

All reported issues were addressed across 8 files (changes from recent commits).

<sub>Tip: Review your code locally with the cubic CLI to iterate faster.

Re-trigger cubic</sub>

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Cursor Bugbot completed as skipped, and this PR exceeds the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cursor]

Cursor Bugbot has reviewed your changes using default effort and found 1 potential issue.

There are 2 total unresolved issues (including 1 from previous review).

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit acfe195. Configure here.}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped with an unresolved finding still open, and this PR exceeds the medium approval threshold. Human review is needed; assigning reviewers.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cubic-dev-ai]

All reported issues were addressed across 72 files (changes from recent commits).

<sub>Tip: Review your code locally with the cubic CLI to iterate faster.

Re-trigger cubic</sub>

[cursor]

Cursor Bugbot has reviewed your changes using default effort and found 1 potential issue.

There are 2 total unresolved issues (including 1 from previous review).

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 0f37e90. Configure here.}

[cursor]

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped with unresolved findings still open, and this PR exceeds the medium approval threshold. Human review is needed; a reviewer is already assigned.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[dastratakos]

Summary

The hosted product (ReviewStage/stage) is sunsetted; this PR vendors its unported functionality into the CLI so users can continue on stagereview. Everything is ported as closely as possible to the hosted implementation (code, prompts, constants, and tests copied verbatim where the architectures align), based on a six-dimension gap analysis of the two repos' code and git histories.

Changes

Chapter generation — prompts (skills/stage-chapters/SKILL.md)

• Full risk-classification rubric with riskLevel/riskReasons outputs (hosted #896)
• Prologue rootCause section (hosted #1051), chapter-level mermaid diagram guidance (hosted #771)
• Restored hosted's full exemplar density, PR title/body + additional-instructions context, keyChange inline-markdown rule

Chapter generation — pipeline & data model (packages/cli, packages/types)

• No-newline markers filtered from parsed hunks (hosted #1132) with regression tests
• Hosted ignore semantics: full DEFAULT_IGNORE_PATTERNS (7 new lockfiles, #871), one case-insensitive matcher so .stageignore! negations re-include defaults
• Zero-hunk files (pure renames) included in Other Changes via the header-only sentinel
• riskLevel/riskReasons + rootCause across types, DB (+migration), ingestion, routes
• What-to-Review ordering by file order then start line (#865/#970); dense 0-based chapterIndex
• prep: PULL REQUEST (title/body, 4k cap), STATS, and ADDITIONAL INSTRUCTIONS sections (.stageinstructions file + --instructions flag, hosted #1036 analog)
• Repair-not-reject line refs; pre-assembled chapter files no longer bypass validation; literal-\n unescaping; consistent diff prefixes

Review UI (packages/web)

• Activity tab: conversation timeline (gh-api-backed), review cards, review-comment diff previews (#1105), events column, read-only reactions
• Continuous chapter review mode (Page/Scroll, #1083) with active-chapter preservation on mode switch
• Risk badges, per-chapter/per-file comment counts, remaining-review-lines sticky header (#1107)
• Markdown stack: shiki highlighting (#796/#826), mermaid in chapter summaries (#771), GitHub image proxy for private attachments (#820)
• Diff extras: image diffs (base64 over the wire), full previews for moved files (#1106), virtualized file list, minimize inline comments
• Small fixes: bot reviewer badges (#1131), resolve-button placement (#1133), sticky-header z-index (#1019), CI popover scroll (#1061), min-height floor (#1014), annotation tinting (#757/#1016), index j/k (#835), dark default theme (#819), derived chapter viewed state (#1021)
• GitHub: stale pending-review recovery (#1072), label management (#1071)
• GitHub viewed-file sync: marks/unmarks files on GitHub per hosted's rules and unions GitHub's viewed state into the local view state (PR runs, best-effort)
• Merge: ruleset-aware allowed merge methods (#1057), single inline merge action with status chip and queue position (#1052)
• Stacked-PR detection and navigation with ⌘⇧↑/⌘⇧↓ (#1049); entries with a local run navigate to it, others link to GitHub
• Review panel keyboard shortcuts: ⌘J toggle, ⌘↩ submit

Configurability

• Diff typography (font/size/line height/ligatures, #974), per-app-theme syntax themes (62), app text size (#1038), chapter panel position (#898), What-to-Review visibility (#1077)

Descoped (depend on hosted backends): Stagent chat, ⌘K action board, reaction toggling, timeline-side comment mutations, mobile layouts, AI chapter editing/regeneration UI.

Testing

• pnpm typecheck, pnpm lint, pnpm test all green (73 files, 643 tests — including regression suites ported from hosted for the diff parser, file filtering, other-changes, line-ref repair, remaining-line counts, timeline assembly, full-file previews, and review recovery)
• pnpm build (SPA + CLI bundle) succeeds

---

Open workspace in Conductor

---

Summary by cubic

Vendors the hosted Stage experience into the CLI so teams can keep using stagereview after the hosted app sunset, with generation parity, a full review UI, GitHub viewed-file sync, labels/stack/merge parity, and configurable display/settings. Also aligns the continuous chapter panel header with paged mode by reusing the shared chapter navigator, with in-page navigation to avoid remounts.

• Bug Fixes

  • Drop stale comment‑draft store writes that settle after stack navigation.
  • Count outdated GitHub review threads via their original coordinates and carry them as a null‑anchored variant so file and chapter comment badges stay accurate.
  • Restore j/k navigation and scroll‑into‑view by typing the --content-top CSS variable, fixing keyboard and deep‑link scrolling in Files and chapter pages.

• Refactors

  • Drop the legacy syntax‑theme localStorage migration; the curated single‑key setting now resets to defaults on upgrade. Remove hosted‑Stage provenance comments (no behavior change).

^{Written for commit 0f37e90. Summary will update on new commits.}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Bugbot skipped with an unresolved finding, Security Agent reported an unresolved high-severity issue, and this PR is above the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit:5d995b7d33

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[cubic-dev-ai]

All reported issues were addressed across 128 files

<sub>Reply with feedback, questions, or to request a fix.

Re-trigger cubic</sub>

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Cursor Bugbot skipped, an unresolved high-severity Security Agent finding remains, unresolved review threads are still open, and this PR exceeds the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cubic-dev-ai]

All reported issues were addressed across 21 files (changes from recent commits).

<sub>Reply with feedback, questions, or to request a fix.

Re-trigger cubic</sub>

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit:570505304f

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Bugbot completed as skipped with an unresolved finding, an unresolved high-severity Security Agent finding remains, and unresolved review threads remain above the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Bugbot skipped with an unresolved finding, an unresolved high-severity Security Agent finding remains, unresolved review threads remain, and this PR exceeds the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cursor]

Cursor Bugbot has reviewed your changes using default effort and found 1 potential issue.

There are 2 total unresolved issues (including 1 from previous review).

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit ba5b0d2. Configure here.}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped with an unresolved high-severity finding still open, and this PR exceeds the medium approval threshold (migrations, GitHub/token-backed routes, lockfile, cross-cutting CLI/web). Human review is needed; assigning reviewers.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit:ba5b0d2b1e

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[cursor]

Cursor Bugbot has reviewed your changes using default effort and found 1 potential issue.

There are 2 total unresolved issues (including 1 from previous review).

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 7b9b3e7. Configure here.}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped with an unresolved finding still open, and this PR exceeds the medium approval threshold (migrations, GitHub/token-backed routes, cross-cutting CLI/web). Human review is needed; assigning reviewers.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cubic-dev-ai]

1 issue found across 5 files (changes from recent commits).

Prompt for AI agents (unresolved issues)

Check if these issues are valid — if so, understand the root cause of each and fix them. If appropriate, use sub-agents to investigate and fix each issue separately.
<file name="packages/cli/src/__tests__/view-state-github-harness.ts">
<violation number="1" location="packages/cli/src/__tests__/view-state-github-harness.ts:130">
P3: The shim hardcodes state OPEN on every pr list response, so the routes tests using this harness cannot exercise the closed/merged-PR fallback in getPullRequestOrThrow (pull-request.ts:196, listParsed.data.find(pr => pr.state === OPEN) ?? listParsed.data[0]). The available options (branchPrNumber or null) only cover the open-PR and no-PR cases, leaving the newest-closed-PR fallback and prefer-open ordering untested by the integration suite. Consider exposing a configurable pr-list entries option on the harness so that fallback gets regression coverage.</violation>
</file>

<sub>Tip: Review your code locally with the cubic CLI to iterate faster.

Re-trigger cubic</sub>

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped, and this PR exceeds the medium approval threshold (migrations, GitHub/token-backed routes, lockfile, cross-cutting CLI/web). Human review is needed; assigning reviewers.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped, and this PR exceeds the medium approval threshold (migrations, GitHub/token-backed routes, lockfile, cross-cutting CLI/web). Human review is needed; assigning reviewers.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cubic-dev-ai]

All reported issues were addressed across 8 files (changes from recent commits).

<sub>Tip: Review your code locally with the cubic CLI to iterate faster.

Re-trigger cubic</sub>

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit:cd06993e8c

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Cursor Bugbot completed as skipped, and this PR exceeds the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cubic-dev-ai]

All reported issues were addressed across 8 files (changes from recent commits).

<sub>Tip: Review your code locally with the cubic CLI to iterate faster.

Re-trigger cubic</sub>

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Cursor Bugbot completed as skipped, and this PR exceeds the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cursor]

Cursor Bugbot has reviewed your changes using default effort and found 1 potential issue.

There are 2 total unresolved issues (including 1 from previous review).

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit acfe195. Configure here.}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped with an unresolved finding still open, and this PR exceeds the medium approval threshold. Human review is needed; assigning reviewers.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cubic-dev-ai]

All reported issues were addressed across 72 files (changes from recent commits).

<sub>Tip: Review your code locally with the cubic CLI to iterate faster.

Re-trigger cubic</sub>

[cursor]

Cursor Bugbot has reviewed your changes using default effort and found 1 potential issue.

There are 2 total unresolved issues (including 1 from previous review).

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 0f37e90. Configure here.}

[cursor]

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped with unresolved findings still open, and this PR exceeds the medium approval threshold. Human review is needed; a reviewer is already assigned.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[dastratakos]

Summary

The hosted product (ReviewStage/stage) is sunsetted; this PR vendors its unported functionality into the CLI so users can continue on stagereview. Everything is ported as closely as possible to the hosted implementation (code, prompts, constants, and tests copied verbatim where the architectures align), based on a six-dimension gap analysis of the two repos' code and git histories.

Changes

Chapter generation — prompts (skills/stage-chapters/SKILL.md)

• Full risk-classification rubric with riskLevel/riskReasons outputs (hosted #896)
• Prologue rootCause section (hosted #1051), chapter-level mermaid diagram guidance (hosted #771)
• Restored hosted's full exemplar density, PR title/body + additional-instructions context, keyChange inline-markdown rule

Chapter generation — pipeline & data model (packages/cli, packages/types)

• No-newline markers filtered from parsed hunks (hosted #1132) with regression tests
• Hosted ignore semantics: full DEFAULT_IGNORE_PATTERNS (7 new lockfiles, #871), one case-insensitive matcher so .stageignore! negations re-include defaults
• Zero-hunk files (pure renames) included in Other Changes via the header-only sentinel
• riskLevel/riskReasons + rootCause across types, DB (+migration), ingestion, routes
• What-to-Review ordering by file order then start line (#865/#970); dense 0-based chapterIndex
• prep: PULL REQUEST (title/body, 4k cap), STATS, and ADDITIONAL INSTRUCTIONS sections (.stageinstructions file + --instructions flag, hosted #1036 analog)
• Repair-not-reject line refs; pre-assembled chapter files no longer bypass validation; literal-\n unescaping; consistent diff prefixes

Review UI (packages/web)

• Activity tab: conversation timeline (gh-api-backed), review cards, review-comment diff previews (#1105), events column, read-only reactions
• Continuous chapter review mode (Page/Scroll, #1083) with active-chapter preservation on mode switch
• Risk badges, per-chapter/per-file comment counts, remaining-review-lines sticky header (#1107)
• Markdown stack: shiki highlighting (#796/#826), mermaid in chapter summaries (#771), GitHub image proxy for private attachments (#820)
• Diff extras: image diffs (base64 over the wire), full previews for moved files (#1106), virtualized file list, minimize inline comments
• Small fixes: bot reviewer badges (#1131), resolve-button placement (#1133), sticky-header z-index (#1019), CI popover scroll (#1061), min-height floor (#1014), annotation tinting (#757/#1016), index j/k (#835), dark default theme (#819), derived chapter viewed state (#1021)
• GitHub: stale pending-review recovery (#1072), label management (#1071)
• GitHub viewed-file sync: marks/unmarks files on GitHub per hosted's rules and unions GitHub's viewed state into the local view state (PR runs, best-effort)
• Merge: ruleset-aware allowed merge methods (#1057), single inline merge action with status chip and queue position (#1052)
• Stacked-PR detection and navigation with ⌘⇧↑/⌘⇧↓ (#1049); entries with a local run navigate to it, others link to GitHub
• Review panel keyboard shortcuts: ⌘J toggle, ⌘↩ submit

Configurability

• Diff typography (font/size/line height/ligatures, #974), per-app-theme syntax themes (62), app text size (#1038), chapter panel position (#898), What-to-Review visibility (#1077)

Descoped (depend on hosted backends): Stagent chat, ⌘K action board, reaction toggling, timeline-side comment mutations, mobile layouts, AI chapter editing/regeneration UI.

Testing

• pnpm typecheck, pnpm lint, pnpm test all green (73 files, 643 tests — including regression suites ported from hosted for the diff parser, file filtering, other-changes, line-ref repair, remaining-line counts, timeline assembly, full-file previews, and review recovery)
• pnpm build (SPA + CLI bundle) succeeds

---

Open workspace in Conductor

---

Summary by cubic

Vendors the hosted Stage experience into the CLI so teams can keep using stagereview after the hosted app sunset, with generation parity, a full review UI, GitHub viewed-file sync, labels/stack/merge parity, and configurable display/settings. Also aligns the continuous chapter panel header with paged mode by reusing the shared chapter navigator, with in-page navigation to avoid remounts.

• Bug Fixes

  • Drop stale comment‑draft store writes that settle after stack navigation.
  • Count outdated GitHub review threads via their original coordinates and carry them as a null‑anchored variant so file and chapter comment badges stay accurate.
  • Restore j/k navigation and scroll‑into‑view by typing the --content-top CSS variable, fixing keyboard and deep‑link scrolling in Files and chapter pages.

• Refactors

  • Drop the legacy syntax‑theme localStorage migration; the curated single‑key setting now resets to defaults on upgrade. Remove hosted‑Stage provenance comments (no behavior change).

^{Written for commit 0f37e90. Summary will update on new commits.}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Bugbot skipped with an unresolved finding, Security Agent reported an unresolved high-severity issue, and this PR is above the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit:5d995b7d33

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[cubic-dev-ai]

All reported issues were addressed across 128 files

<sub>Reply with feedback, questions, or to request a fix.

Re-trigger cubic</sub>

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Cursor Bugbot skipped, an unresolved high-severity Security Agent finding remains, unresolved review threads are still open, and this PR exceeds the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cubic-dev-ai]

All reported issues were addressed across 21 files (changes from recent commits).

<sub>Reply with feedback, questions, or to request a fix.

Re-trigger cubic</sub>

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit:570505304f

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Bugbot completed as skipped with an unresolved finding, an unresolved high-severity Security Agent finding remains, and unresolved review threads remain above the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Bugbot skipped with an unresolved finding, an unresolved high-severity Security Agent finding remains, unresolved review threads remain, and this PR exceeds the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cursor]

Cursor Bugbot has reviewed your changes using default effort and found 1 potential issue.

There are 2 total unresolved issues (including 1 from previous review).

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit ba5b0d2. Configure here.}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped with an unresolved high-severity finding still open, and this PR exceeds the medium approval threshold (migrations, GitHub/token-backed routes, lockfile, cross-cutting CLI/web). Human review is needed; assigning reviewers.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit:ba5b0d2b1e

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[cursor]

Cursor Bugbot has reviewed your changes using default effort and found 1 potential issue.

There are 2 total unresolved issues (including 1 from previous review).

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 7b9b3e7. Configure here.}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped with an unresolved finding still open, and this PR exceeds the medium approval threshold (migrations, GitHub/token-backed routes, cross-cutting CLI/web). Human review is needed; assigning reviewers.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cubic-dev-ai]

1 issue found across 5 files (changes from recent commits).

Prompt for AI agents (unresolved issues)

Check if these issues are valid — if so, understand the root cause of each and fix them. If appropriate, use sub-agents to investigate and fix each issue separately.
<file name="packages/cli/src/__tests__/view-state-github-harness.ts">
<violation number="1" location="packages/cli/src/__tests__/view-state-github-harness.ts:130">
P3: The shim hardcodes state OPEN on every pr list response, so the routes tests using this harness cannot exercise the closed/merged-PR fallback in getPullRequestOrThrow (pull-request.ts:196, listParsed.data.find(pr => pr.state === OPEN) ?? listParsed.data[0]). The available options (branchPrNumber or null) only cover the open-PR and no-PR cases, leaving the newest-closed-PR fallback and prefer-open ordering untested by the integration suite. Consider exposing a configurable pr-list entries option on the harness so that fallback gets regression coverage.</violation>
</file>

<sub>Tip: Review your code locally with the cubic CLI to iterate faster.

Re-trigger cubic</sub>

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped, and this PR exceeds the medium approval threshold (migrations, GitHub/token-backed routes, lockfile, cross-cutting CLI/web). Human review is needed; assigning reviewers.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped, and this PR exceeds the medium approval threshold (migrations, GitHub/token-backed routes, lockfile, cross-cutting CLI/web). Human review is needed; assigning reviewers.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cubic-dev-ai]

All reported issues were addressed across 8 files (changes from recent commits).

<sub>Tip: Review your code locally with the cubic CLI to iterate faster.

Re-trigger cubic</sub>

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit:cd06993e8c

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Cursor Bugbot completed as skipped, and this PR exceeds the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cubic-dev-ai]

All reported issues were addressed across 8 files (changes from recent commits).

<sub>Tip: Review your code locally with the cubic CLI to iterate faster.

Re-trigger cubic</sub>

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Cursor Bugbot completed as skipped, and this PR exceeds the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cursor]

Cursor Bugbot has reviewed your changes using default effort and found 1 potential issue.

There are 2 total unresolved issues (including 1 from previous review).

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit acfe195. Configure here.}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped with an unresolved finding still open, and this PR exceeds the medium approval threshold. Human review is needed; assigning reviewers.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cubic-dev-ai]

All reported issues were addressed across 72 files (changes from recent commits).

<sub>Tip: Review your code locally with the cubic CLI to iterate faster.

Re-trigger cubic</sub>

[cursor]

Cursor Bugbot has reviewed your changes using default effort and found 1 potential issue.

There are 2 total unresolved issues (including 1 from previous review).

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 0f37e90. Configure here.}

[cursor]

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped with unresolved findings still open, and this PR exceeds the medium approval threshold. Human review is needed; a reviewer is already assigned.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[dastratakos]

Summary

The hosted product (ReviewStage/stage) is sunsetted; this PR vendors its unported functionality into the CLI so users can continue on stagereview. Everything is ported as closely as possible to the hosted implementation (code, prompts, constants, and tests copied verbatim where the architectures align), based on a six-dimension gap analysis of the two repos' code and git histories.

Changes

Chapter generation — prompts (skills/stage-chapters/SKILL.md)

• Full risk-classification rubric with riskLevel/riskReasons outputs (hosted #896)
• Prologue rootCause section (hosted #1051), chapter-level mermaid diagram guidance (hosted #771)
• Restored hosted's full exemplar density, PR title/body + additional-instructions context, keyChange inline-markdown rule

Chapter generation — pipeline & data model (packages/cli, packages/types)

• No-newline markers filtered from parsed hunks (hosted #1132) with regression tests
• Hosted ignore semantics: full DEFAULT_IGNORE_PATTERNS (7 new lockfiles, #871), one case-insensitive matcher so .stageignore! negations re-include defaults
• Zero-hunk files (pure renames) included in Other Changes via the header-only sentinel
• riskLevel/riskReasons + rootCause across types, DB (+migration), ingestion, routes
• What-to-Review ordering by file order then start line (#865/#970); dense 0-based chapterIndex
• prep: PULL REQUEST (title/body, 4k cap), STATS, and ADDITIONAL INSTRUCTIONS sections (.stageinstructions file + --instructions flag, hosted #1036 analog)
• Repair-not-reject line refs; pre-assembled chapter files no longer bypass validation; literal-\n unescaping; consistent diff prefixes

Review UI (packages/web)

• Activity tab: conversation timeline (gh-api-backed), review cards, review-comment diff previews (#1105), events column, read-only reactions
• Continuous chapter review mode (Page/Scroll, #1083) with active-chapter preservation on mode switch
• Risk badges, per-chapter/per-file comment counts, remaining-review-lines sticky header (#1107)
• Markdown stack: shiki highlighting (#796/#826), mermaid in chapter summaries (#771), GitHub image proxy for private attachments (#820)
• Diff extras: image diffs (base64 over the wire), full previews for moved files (#1106), virtualized file list, minimize inline comments
• Small fixes: bot reviewer badges (#1131), resolve-button placement (#1133), sticky-header z-index (#1019), CI popover scroll (#1061), min-height floor (#1014), annotation tinting (#757/#1016), index j/k (#835), dark default theme (#819), derived chapter viewed state (#1021)
• GitHub: stale pending-review recovery (#1072), label management (#1071)
• GitHub viewed-file sync: marks/unmarks files on GitHub per hosted's rules and unions GitHub's viewed state into the local view state (PR runs, best-effort)
• Merge: ruleset-aware allowed merge methods (#1057), single inline merge action with status chip and queue position (#1052)
• Stacked-PR detection and navigation with ⌘⇧↑/⌘⇧↓ (#1049); entries with a local run navigate to it, others link to GitHub
• Review panel keyboard shortcuts: ⌘J toggle, ⌘↩ submit

Configurability

• Diff typography (font/size/line height/ligatures, #974), per-app-theme syntax themes (62), app text size (#1038), chapter panel position (#898), What-to-Review visibility (#1077)

Descoped (depend on hosted backends): Stagent chat, ⌘K action board, reaction toggling, timeline-side comment mutations, mobile layouts, AI chapter editing/regeneration UI.

Testing

• pnpm typecheck, pnpm lint, pnpm test all green (73 files, 643 tests — including regression suites ported from hosted for the diff parser, file filtering, other-changes, line-ref repair, remaining-line counts, timeline assembly, full-file previews, and review recovery)
• pnpm build (SPA + CLI bundle) succeeds

---

Open workspace in Conductor

---

Summary by cubic

Vendors the hosted Stage experience into the CLI so teams can keep using stagereview after the hosted app sunset, with generation parity, a full review UI, GitHub viewed-file sync, labels/stack/merge parity, and configurable display/settings. Also aligns the continuous chapter panel header with paged mode by reusing the shared chapter navigator, with in-page navigation to avoid remounts.

• Bug Fixes

  • Drop stale comment‑draft store writes that settle after stack navigation.
  • Count outdated GitHub review threads via their original coordinates and carry them as a null‑anchored variant so file and chapter comment badges stay accurate.
  • Restore j/k navigation and scroll‑into‑view by typing the --content-top CSS variable, fixing keyboard and deep‑link scrolling in Files and chapter pages.

• Refactors

  • Drop the legacy syntax‑theme localStorage migration; the curated single‑key setting now resets to defaults on upgrade. Remove hosted‑Stage provenance comments (no behavior change).

^{Written for commit 0f37e90. Summary will update on new commits.}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Bugbot skipped with an unresolved finding, Security Agent reported an unresolved high-severity issue, and this PR is above the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit:5d995b7d33

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[cubic-dev-ai]

All reported issues were addressed across 128 files

<sub>Reply with feedback, questions, or to request a fix.

Re-trigger cubic</sub>

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Cursor Bugbot skipped, an unresolved high-severity Security Agent finding remains, unresolved review threads are still open, and this PR exceeds the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cubic-dev-ai]

All reported issues were addressed across 21 files (changes from recent commits).

<sub>Reply with feedback, questions, or to request a fix.

Re-trigger cubic</sub>

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit:570505304f

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Bugbot completed as skipped with an unresolved finding, an unresolved high-severity Security Agent finding remains, and unresolved review threads remain above the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Bugbot skipped with an unresolved finding, an unresolved high-severity Security Agent finding remains, unresolved review threads remain, and this PR exceeds the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cursor]

Cursor Bugbot has reviewed your changes using default effort and found 1 potential issue.

There are 2 total unresolved issues (including 1 from previous review).

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit ba5b0d2. Configure here.}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped with an unresolved high-severity finding still open, and this PR exceeds the medium approval threshold (migrations, GitHub/token-backed routes, lockfile, cross-cutting CLI/web). Human review is needed; assigning reviewers.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit:ba5b0d2b1e

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[cursor]

Cursor Bugbot has reviewed your changes using default effort and found 1 potential issue.

There are 2 total unresolved issues (including 1 from previous review).

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 7b9b3e7. Configure here.}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped with an unresolved finding still open, and this PR exceeds the medium approval threshold (migrations, GitHub/token-backed routes, cross-cutting CLI/web). Human review is needed; assigning reviewers.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cubic-dev-ai]

1 issue found across 5 files (changes from recent commits).

Prompt for AI agents (unresolved issues)

Check if these issues are valid — if so, understand the root cause of each and fix them. If appropriate, use sub-agents to investigate and fix each issue separately.
<file name="packages/cli/src/__tests__/view-state-github-harness.ts">
<violation number="1" location="packages/cli/src/__tests__/view-state-github-harness.ts:130">
P3: The shim hardcodes state OPEN on every pr list response, so the routes tests using this harness cannot exercise the closed/merged-PR fallback in getPullRequestOrThrow (pull-request.ts:196, listParsed.data.find(pr => pr.state === OPEN) ?? listParsed.data[0]). The available options (branchPrNumber or null) only cover the open-PR and no-PR cases, leaving the newest-closed-PR fallback and prefer-open ordering untested by the integration suite. Consider exposing a configurable pr-list entries option on the harness so that fallback gets regression coverage.</violation>
</file>

<sub>Tip: Review your code locally with the cubic CLI to iterate faster.

Re-trigger cubic</sub>

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped, and this PR exceeds the medium approval threshold (migrations, GitHub/token-backed routes, lockfile, cross-cutting CLI/web). Human review is needed; assigning reviewers.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped, and this PR exceeds the medium approval threshold (migrations, GitHub/token-backed routes, lockfile, cross-cutting CLI/web). Human review is needed; assigning reviewers.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cubic-dev-ai]

All reported issues were addressed across 8 files (changes from recent commits).

<sub>Tip: Review your code locally with the cubic CLI to iterate faster.

Re-trigger cubic</sub>

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit:cd06993e8c

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Cursor Bugbot completed as skipped, and this PR exceeds the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cubic-dev-ai]

All reported issues were addressed across 8 files (changes from recent commits).

<sub>Tip: Review your code locally with the cubic CLI to iterate faster.

Re-trigger cubic</sub>

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Cursor Bugbot completed as skipped, and this PR exceeds the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cursor]

Cursor Bugbot has reviewed your changes using default effort and found 1 potential issue.

There are 2 total unresolved issues (including 1 from previous review).

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit acfe195. Configure here.}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped with an unresolved finding still open, and this PR exceeds the medium approval threshold. Human review is needed; assigning reviewers.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cubic-dev-ai]

All reported issues were addressed across 72 files (changes from recent commits).

<sub>Tip: Review your code locally with the cubic CLI to iterate faster.

Re-trigger cubic</sub>

[cursor]

Cursor Bugbot has reviewed your changes using default effort and found 1 potential issue.

There are 2 total unresolved issues (including 1 from previous review).

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 0f37e90. Configure here.}

[cursor]

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped with unresolved findings still open, and this PR exceeds the medium approval threshold. Human review is needed; a reviewer is already assigned.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[dastratakos]

Summary

The hosted product (ReviewStage/stage) is sunsetted; this PR vendors its unported functionality into the CLI so users can continue on stagereview. Everything is ported as closely as possible to the hosted implementation (code, prompts, constants, and tests copied verbatim where the architectures align), based on a six-dimension gap analysis of the two repos' code and git histories.

Changes

Chapter generation — prompts (skills/stage-chapters/SKILL.md)

• Full risk-classification rubric with riskLevel/riskReasons outputs (hosted #896)
• Prologue rootCause section (hosted #1051), chapter-level mermaid diagram guidance (hosted #771)
• Restored hosted's full exemplar density, PR title/body + additional-instructions context, keyChange inline-markdown rule

Chapter generation — pipeline & data model (packages/cli, packages/types)

• No-newline markers filtered from parsed hunks (hosted #1132) with regression tests
• Hosted ignore semantics: full DEFAULT_IGNORE_PATTERNS (7 new lockfiles, #871), one case-insensitive matcher so .stageignore! negations re-include defaults
• Zero-hunk files (pure renames) included in Other Changes via the header-only sentinel
• riskLevel/riskReasons + rootCause across types, DB (+migration), ingestion, routes
• What-to-Review ordering by file order then start line (#865/#970); dense 0-based chapterIndex
• prep: PULL REQUEST (title/body, 4k cap), STATS, and ADDITIONAL INSTRUCTIONS sections (.stageinstructions file + --instructions flag, hosted #1036 analog)
• Repair-not-reject line refs; pre-assembled chapter files no longer bypass validation; literal-\n unescaping; consistent diff prefixes

Review UI (packages/web)

• Activity tab: conversation timeline (gh-api-backed), review cards, review-comment diff previews (#1105), events column, read-only reactions
• Continuous chapter review mode (Page/Scroll, #1083) with active-chapter preservation on mode switch
• Risk badges, per-chapter/per-file comment counts, remaining-review-lines sticky header (#1107)
• Markdown stack: shiki highlighting (#796/#826), mermaid in chapter summaries (#771), GitHub image proxy for private attachments (#820)
• Diff extras: image diffs (base64 over the wire), full previews for moved files (#1106), virtualized file list, minimize inline comments
• Small fixes: bot reviewer badges (#1131), resolve-button placement (#1133), sticky-header z-index (#1019), CI popover scroll (#1061), min-height floor (#1014), annotation tinting (#757/#1016), index j/k (#835), dark default theme (#819), derived chapter viewed state (#1021)
• GitHub: stale pending-review recovery (#1072), label management (#1071)
• GitHub viewed-file sync: marks/unmarks files on GitHub per hosted's rules and unions GitHub's viewed state into the local view state (PR runs, best-effort)
• Merge: ruleset-aware allowed merge methods (#1057), single inline merge action with status chip and queue position (#1052)
• Stacked-PR detection and navigation with ⌘⇧↑/⌘⇧↓ (#1049); entries with a local run navigate to it, others link to GitHub
• Review panel keyboard shortcuts: ⌘J toggle, ⌘↩ submit

Configurability

• Diff typography (font/size/line height/ligatures, #974), per-app-theme syntax themes (62), app text size (#1038), chapter panel position (#898), What-to-Review visibility (#1077)

Descoped (depend on hosted backends): Stagent chat, ⌘K action board, reaction toggling, timeline-side comment mutations, mobile layouts, AI chapter editing/regeneration UI.

Testing

• pnpm typecheck, pnpm lint, pnpm test all green (73 files, 643 tests — including regression suites ported from hosted for the diff parser, file filtering, other-changes, line-ref repair, remaining-line counts, timeline assembly, full-file previews, and review recovery)
• pnpm build (SPA + CLI bundle) succeeds

---

Open workspace in Conductor

---

Summary by cubic

Vendors the hosted Stage experience into the CLI so teams can keep using stagereview after the hosted app sunset, with generation parity, a full review UI, GitHub viewed-file sync, labels/stack/merge parity, and configurable display/settings. Also aligns the continuous chapter panel header with paged mode by reusing the shared chapter navigator, with in-page navigation to avoid remounts.

• Bug Fixes

  • Drop stale comment‑draft store writes that settle after stack navigation.
  • Count outdated GitHub review threads via their original coordinates and carry them as a null‑anchored variant so file and chapter comment badges stay accurate.
  • Restore j/k navigation and scroll‑into‑view by typing the --content-top CSS variable, fixing keyboard and deep‑link scrolling in Files and chapter pages.

• Refactors

  • Drop the legacy syntax‑theme localStorage migration; the curated single‑key setting now resets to defaults on upgrade. Remove hosted‑Stage provenance comments (no behavior change).

^{Written for commit 0f37e90. Summary will update on new commits.}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Bugbot skipped with an unresolved finding, Security Agent reported an unresolved high-severity issue, and this PR is above the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit:5d995b7d33

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[cubic-dev-ai]

All reported issues were addressed across 128 files

<sub>Reply with feedback, questions, or to request a fix.

Re-trigger cubic</sub>

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Cursor Bugbot skipped, an unresolved high-severity Security Agent finding remains, unresolved review threads are still open, and this PR exceeds the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cubic-dev-ai]

All reported issues were addressed across 21 files (changes from recent commits).

<sub>Reply with feedback, questions, or to request a fix.

Re-trigger cubic</sub>

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit:570505304f

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Bugbot completed as skipped with an unresolved finding, an unresolved high-severity Security Agent finding remains, and unresolved review threads remain above the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Bugbot skipped with an unresolved finding, an unresolved high-severity Security Agent finding remains, unresolved review threads remain, and this PR exceeds the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cursor]

Cursor Bugbot has reviewed your changes using default effort and found 1 potential issue.

There are 2 total unresolved issues (including 1 from previous review).

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit ba5b0d2. Configure here.}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped with an unresolved high-severity finding still open, and this PR exceeds the medium approval threshold (migrations, GitHub/token-backed routes, lockfile, cross-cutting CLI/web). Human review is needed; assigning reviewers.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit:ba5b0d2b1e

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[cursor]

Cursor Bugbot has reviewed your changes using default effort and found 1 potential issue.

There are 2 total unresolved issues (including 1 from previous review).

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 7b9b3e7. Configure here.}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped with an unresolved finding still open, and this PR exceeds the medium approval threshold (migrations, GitHub/token-backed routes, cross-cutting CLI/web). Human review is needed; assigning reviewers.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cubic-dev-ai]

1 issue found across 5 files (changes from recent commits).

Prompt for AI agents (unresolved issues)

Check if these issues are valid — if so, understand the root cause of each and fix them. If appropriate, use sub-agents to investigate and fix each issue separately.
<file name="packages/cli/src/__tests__/view-state-github-harness.ts">
<violation number="1" location="packages/cli/src/__tests__/view-state-github-harness.ts:130">
P3: The shim hardcodes state OPEN on every pr list response, so the routes tests using this harness cannot exercise the closed/merged-PR fallback in getPullRequestOrThrow (pull-request.ts:196, listParsed.data.find(pr => pr.state === OPEN) ?? listParsed.data[0]). The available options (branchPrNumber or null) only cover the open-PR and no-PR cases, leaving the newest-closed-PR fallback and prefer-open ordering untested by the integration suite. Consider exposing a configurable pr-list entries option on the harness so that fallback gets regression coverage.</violation>
</file>

<sub>Tip: Review your code locally with the cubic CLI to iterate faster.

Re-trigger cubic</sub>

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped, and this PR exceeds the medium approval threshold (migrations, GitHub/token-backed routes, lockfile, cross-cutting CLI/web). Human review is needed; assigning reviewers.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped, and this PR exceeds the medium approval threshold (migrations, GitHub/token-backed routes, lockfile, cross-cutting CLI/web). Human review is needed; assigning reviewers.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cubic-dev-ai]

All reported issues were addressed across 8 files (changes from recent commits).

<sub>Tip: Review your code locally with the cubic CLI to iterate faster.

Re-trigger cubic</sub>

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit:cd06993e8c

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Cursor Bugbot completed as skipped, and this PR exceeds the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cubic-dev-ai]

All reported issues were addressed across 8 files (changes from recent commits).

<sub>Tip: Review your code locally with the cubic CLI to iterate faster.

Re-trigger cubic</sub>

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Cursor Bugbot completed as skipped, and this PR exceeds the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cursor]

Cursor Bugbot has reviewed your changes using default effort and found 1 potential issue.

There are 2 total unresolved issues (including 1 from previous review).

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit acfe195. Configure here.}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped with an unresolved finding still open, and this PR exceeds the medium approval threshold. Human review is needed; assigning reviewers.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cubic-dev-ai]

All reported issues were addressed across 72 files (changes from recent commits).

<sub>Tip: Review your code locally with the cubic CLI to iterate faster.

Re-trigger cubic</sub>

[cursor]

Cursor Bugbot has reviewed your changes using default effort and found 1 potential issue.

There are 2 total unresolved issues (including 1 from previous review).

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 0f37e90. Configure here.}

[cursor]

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped with unresolved findings still open, and this PR exceeds the medium approval threshold. Human review is needed; a reviewer is already assigned.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[dastratakos]

Summary

The hosted product (ReviewStage/stage) is sunsetted; this PR vendors its unported functionality into the CLI so users can continue on stagereview. Everything is ported as closely as possible to the hosted implementation (code, prompts, constants, and tests copied verbatim where the architectures align), based on a six-dimension gap analysis of the two repos' code and git histories.

Changes

Chapter generation — prompts (skills/stage-chapters/SKILL.md)

• Full risk-classification rubric with riskLevel/riskReasons outputs (hosted #896)
• Prologue rootCause section (hosted #1051), chapter-level mermaid diagram guidance (hosted #771)
• Restored hosted's full exemplar density, PR title/body + additional-instructions context, keyChange inline-markdown rule

Chapter generation — pipeline & data model (packages/cli, packages/types)

• No-newline markers filtered from parsed hunks (hosted #1132) with regression tests
• Hosted ignore semantics: full DEFAULT_IGNORE_PATTERNS (7 new lockfiles, #871), one case-insensitive matcher so .stageignore! negations re-include defaults
• Zero-hunk files (pure renames) included in Other Changes via the header-only sentinel
• riskLevel/riskReasons + rootCause across types, DB (+migration), ingestion, routes
• What-to-Review ordering by file order then start line (#865/#970); dense 0-based chapterIndex
• prep: PULL REQUEST (title/body, 4k cap), STATS, and ADDITIONAL INSTRUCTIONS sections (.stageinstructions file + --instructions flag, hosted #1036 analog)
• Repair-not-reject line refs; pre-assembled chapter files no longer bypass validation; literal-\n unescaping; consistent diff prefixes

Review UI (packages/web)

• Activity tab: conversation timeline (gh-api-backed), review cards, review-comment diff previews (#1105), events column, read-only reactions
• Continuous chapter review mode (Page/Scroll, #1083) with active-chapter preservation on mode switch
• Risk badges, per-chapter/per-file comment counts, remaining-review-lines sticky header (#1107)
• Markdown stack: shiki highlighting (#796/#826), mermaid in chapter summaries (#771), GitHub image proxy for private attachments (#820)
• Diff extras: image diffs (base64 over the wire), full previews for moved files (#1106), virtualized file list, minimize inline comments
• Small fixes: bot reviewer badges (#1131), resolve-button placement (#1133), sticky-header z-index (#1019), CI popover scroll (#1061), min-height floor (#1014), annotation tinting (#757/#1016), index j/k (#835), dark default theme (#819), derived chapter viewed state (#1021)
• GitHub: stale pending-review recovery (#1072), label management (#1071)
• GitHub viewed-file sync: marks/unmarks files on GitHub per hosted's rules and unions GitHub's viewed state into the local view state (PR runs, best-effort)
• Merge: ruleset-aware allowed merge methods (#1057), single inline merge action with status chip and queue position (#1052)
• Stacked-PR detection and navigation with ⌘⇧↑/⌘⇧↓ (#1049); entries with a local run navigate to it, others link to GitHub
• Review panel keyboard shortcuts: ⌘J toggle, ⌘↩ submit

Configurability

• Diff typography (font/size/line height/ligatures, #974), per-app-theme syntax themes (62), app text size (#1038), chapter panel position (#898), What-to-Review visibility (#1077)

Descoped (depend on hosted backends): Stagent chat, ⌘K action board, reaction toggling, timeline-side comment mutations, mobile layouts, AI chapter editing/regeneration UI.

Testing

• pnpm typecheck, pnpm lint, pnpm test all green (73 files, 643 tests — including regression suites ported from hosted for the diff parser, file filtering, other-changes, line-ref repair, remaining-line counts, timeline assembly, full-file previews, and review recovery)
• pnpm build (SPA + CLI bundle) succeeds

---

Open workspace in Conductor

---

Summary by cubic

Vendors the hosted Stage experience into the CLI so teams can keep using stagereview after the hosted app sunset, with generation parity, a full review UI, GitHub viewed-file sync, labels/stack/merge parity, and configurable display/settings. Also aligns the continuous chapter panel header with paged mode by reusing the shared chapter navigator, with in-page navigation to avoid remounts.

• Bug Fixes

  • Drop stale comment‑draft store writes that settle after stack navigation.
  • Count outdated GitHub review threads via their original coordinates and carry them as a null‑anchored variant so file and chapter comment badges stay accurate.
  • Restore j/k navigation and scroll‑into‑view by typing the --content-top CSS variable, fixing keyboard and deep‑link scrolling in Files and chapter pages.

• Refactors

  • Drop the legacy syntax‑theme localStorage migration; the curated single‑key setting now resets to defaults on upgrade. Remove hosted‑Stage provenance comments (no behavior change).

^{Written for commit 0f37e90. Summary will update on new commits.}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Bugbot skipped with an unresolved finding, Security Agent reported an unresolved high-severity issue, and this PR is above the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit:5d995b7d33

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[cubic-dev-ai]

All reported issues were addressed across 128 files

<sub>Reply with feedback, questions, or to request a fix.

Re-trigger cubic</sub>

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Cursor Bugbot skipped, an unresolved high-severity Security Agent finding remains, unresolved review threads are still open, and this PR exceeds the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cubic-dev-ai]

All reported issues were addressed across 21 files (changes from recent commits).

<sub>Reply with feedback, questions, or to request a fix.

Re-trigger cubic</sub>

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit:570505304f

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Bugbot completed as skipped with an unresolved finding, an unresolved high-severity Security Agent finding remains, and unresolved review threads remain above the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Bugbot skipped with an unresolved finding, an unresolved high-severity Security Agent finding remains, unresolved review threads remain, and this PR exceeds the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cursor]

Cursor Bugbot has reviewed your changes using default effort and found 1 potential issue.

There are 2 total unresolved issues (including 1 from previous review).

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit ba5b0d2. Configure here.}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped with an unresolved high-severity finding still open, and this PR exceeds the medium approval threshold (migrations, GitHub/token-backed routes, lockfile, cross-cutting CLI/web). Human review is needed; assigning reviewers.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit:ba5b0d2b1e

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[cursor]

Cursor Bugbot has reviewed your changes using default effort and found 1 potential issue.

There are 2 total unresolved issues (including 1 from previous review).

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 7b9b3e7. Configure here.}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped with an unresolved finding still open, and this PR exceeds the medium approval threshold (migrations, GitHub/token-backed routes, cross-cutting CLI/web). Human review is needed; assigning reviewers.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cubic-dev-ai]

1 issue found across 5 files (changes from recent commits).

Prompt for AI agents (unresolved issues)

Check if these issues are valid — if so, understand the root cause of each and fix them. If appropriate, use sub-agents to investigate and fix each issue separately.
<file name="packages/cli/src/__tests__/view-state-github-harness.ts">
<violation number="1" location="packages/cli/src/__tests__/view-state-github-harness.ts:130">
P3: The shim hardcodes state OPEN on every pr list response, so the routes tests using this harness cannot exercise the closed/merged-PR fallback in getPullRequestOrThrow (pull-request.ts:196, listParsed.data.find(pr => pr.state === OPEN) ?? listParsed.data[0]). The available options (branchPrNumber or null) only cover the open-PR and no-PR cases, leaving the newest-closed-PR fallback and prefer-open ordering untested by the integration suite. Consider exposing a configurable pr-list entries option on the harness so that fallback gets regression coverage.</violation>
</file>

<sub>Tip: Review your code locally with the cubic CLI to iterate faster.

Re-trigger cubic</sub>

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped, and this PR exceeds the medium approval threshold (migrations, GitHub/token-backed routes, lockfile, cross-cutting CLI/web). Human review is needed; assigning reviewers.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped, and this PR exceeds the medium approval threshold (migrations, GitHub/token-backed routes, lockfile, cross-cutting CLI/web). Human review is needed; assigning reviewers.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cubic-dev-ai]

All reported issues were addressed across 8 files (changes from recent commits).

<sub>Tip: Review your code locally with the cubic CLI to iterate faster.

Re-trigger cubic</sub>

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit:cd06993e8c

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Cursor Bugbot completed as skipped, and this PR exceeds the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cubic-dev-ai]

All reported issues were addressed across 8 files (changes from recent commits).

<sub>Tip: Review your code locally with the cubic CLI to iterate faster.

Re-trigger cubic</sub>

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and assigned reviewers. Not approving: Cursor Bugbot completed as skipped, and this PR exceeds the medium risk threshold.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cursor]

Cursor Bugbot has reviewed your changes using default effort and found 1 potential issue.

There are 2 total unresolved issues (including 1 from previous review).

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit acfe195. Configure here.}

[cursor]

Stale comment

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped with an unresolved finding still open, and this PR exceeds the medium approval threshold. Human review is needed; assigning reviewers.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}

[cubic-dev-ai]

All reported issues were addressed across 72 files (changes from recent commits).

<sub>Tip: Review your code locally with the cubic CLI to iterate faster.

Re-trigger cubic</sub>

[cursor]

Cursor Bugbot has reviewed your changes using default effort and found 1 potential issue.

There are 2 total unresolved issues (including 1 from previous review).

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 0f37e90. Configure here.}

[cursor]

Risk: high. Left a non-blocking comment and did not approve: Cursor Bugbot completed as skipped with unresolved findings still open, and this PR exceeds the medium approval threshold. Human review is needed; a reviewer is already assigned.

_{Sent by Cursor Approval Agent: Pull Request Router and Approver}