## Summary
- Improve compaction prompt/provider option handling in IPC handlers
- Tighten local agent tool option passing and argument defaults
- Align LLM engine provider option generation across related paths

## Test plan
- npm run fmt
- npm run lint:fix
- npm run ts

🤖 Generated with [Claude Code](https://claude.com/claude-code)
<!-- devin-review-badge-begin -->

---

<a href="https://app.devin.ai/review/dyad-sh/dyad/pull/2876"
target="_blank">
  <picture>
<source media="(prefers-color-scheme: dark)"
srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img
src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1"
alt="Open with Devin">
  </picture>
</a>
<!-- devin-review-badge-end -->

## Summary
- Update the Claude triage workflow to be more conservative about
changing user-provided issue titles
- Only update titles that are essentially blank or contain no meaningful
information (placeholders, single unhelpful words like "help" or "bug",
just punctuation)
- Key rule: if you can understand what the user is roughly talking about
from the title, leave it alone

## Test plan
- Review the updated workflow conditions to verify they are
appropriately conservative
- Monitor triage behavior on new issues to confirm fewer unnecessary
title edits

🤖 Generated with [Claude Code](https://claude.com/claude-code)
<!-- devin-review-badge-begin -->

---

<a href="https://app.devin.ai/review/dyad-sh/dyad/pull/2888"
target="_blank">
  <picture>
<source media="(prefers-color-scheme: dark)"
srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img
src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1"
alt="Open with Devin">
  </picture>
</a>
<!-- devin-review-badge-end -->
Co-authored-by: Will Chen <willchen90@gmail.com>
Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com>

## Summary
- Include `dyadRequestId` in `terminated_stream_retry` telemetry events
for better traceability
- Add new `terminated_stream_retries_exhausted` telemetry event when all
retries are used up (both stream_iteration and response_finalization
phases)
- Bump `MAX_TERMINATED_STREAM_RETRIES` from 2 to 3

## Test plan
- Existing `local_agent_handler` unit tests pass (stream retry tests
verified)
- Verify telemetry events include `dyadRequestId` in retry scenarios
- Verify exhausted event fires when retries exceed the new limit of 3

#skip-bugbot

Made with [Cursor](https://cursor.com)
<!-- devin-review-badge-begin -->

---

<a href="https://app.devin.ai/review/dyad-sh/dyad/pull/2877"
target="_blank">
  <picture>
<source media="(prefers-color-scheme: dark)"
srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img
src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1"
alt="Open with Devin">
  </picture>
</a>
<!-- devin-review-badge-end -->

## Summary
- Add `wwwillchen-bot` to `allowed_non_write_users` in
`.github/workflows/claude-pr-review.yml`
- Keep existing `princeaden1` allowlist entry and broaden workflow
access for trusted bot-driven contributions

## Test plan
- npm run fmt
- npm run lint:fix
- npm run ts

#skip-bugbot

🤖 Generated with [Claude Code](https://claude.com/claude-code)
<!-- devin-review-badge-begin -->

---

<a href="https://app.devin.ai/review/dyad-sh/dyad/pull/2882"
target="_blank">
  <picture>
<source media="(prefers-color-scheme: dark)"
srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img
src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1"
alt="Open with Devin">
  </picture>
</a>
<!-- devin-review-badge-end -->

## Summary
- Increase MAX_TOOL_CALL_STEPS from 25 to 50 to allow longer multi-step
tasks
- Add step limit detection to track total steps across all passes
- Show a clear `<dyad-step-limit>` message when the limit is reached,
instructing users to send "continue" to resume
- Create DyadStepLimit component for displaying the pause notification

Fixes #2754

## Test plan
- Run the local agent and perform a task that requires many tool calls
- Verify the agent pauses at 50 tool calls instead of 25
- Verify a clear message is shown explaining why it paused and how to
continue
- Type "continue" to verify the agent resumes working

🤖 Generated with [Claude Code](https://claude.com/claude-code)
<!-- devin-review-badge-begin -->

---

<a href="https://app.devin.ai/review/dyad-sh/dyad/pull/2828"
target="_blank">
  <picture>
<source media="(prefers-color-scheme: dark)"
srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img
src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1"
alt="Open with Devin">
  </picture>
</a>
<!-- devin-review-badge-end -->

---------
Co-authored-by: Will Chen <willchen90@gmail.com>
Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com>

## Summary
- Add a pre-flight check in `gitAdd` to detect if a file is ignored by
`.gitignore` before attempting to stage it
- Skip staging ignored files gracefully with a debug log instead of
throwing an error
- This prevents errors when the AI attempts to stage files like
`.env.local` that should remain untracked

## Test plan
- [ ] Verify that staging a normal file still works correctly
- [ ] Create a `.gitignore` file that ignores a test file (e.g.,
`test.ignored`)
- [ ] Create the ignored file and attempt to stage it
- [ ] Confirm the file is skipped without error and a debug message is
logged

🤖 Generated with [Claude Code](https://claude.com/claude-code)
<!-- devin-review-badge-begin -->

---

<a href="https://app.devin.ai/review/dyad-sh/dyad/pull/2707"
target="_blank">
  <picture>
<source media="(prefers-color-scheme: dark)"
srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img
src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1"
alt="Open with Devin">
  </picture>
</a>
<!-- devin-review-badge-end -->

---------
Co-authored-by: Will Chen <willchen90@gmail.com>
Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com>
Co-authored-by: claude[bot] <41898282+claude[bot]@users.noreply.github.com>

## Summary
- Fix stale-read race condition in `handleSupabaseOAuthReturn` and
`refreshSupabaseTokenForOrganization` where settings are read before an
async network call, then written back using the stale snapshot —
silently overwriting any concurrent credential writes from other orgs
- Fix legacy `refreshSupabaseToken` to preserve the `organizations` map
by spreading existing supabase settings before writing

## Test plan
- Verify connecting multiple Supabase organizations no longer causes
previously-connected orgs to lose their credentials
- Verify token refresh for one org does not clobber another org's
credentials
- All existing unit tests pass (842/842)

🤖 Generated with [Claude Code](https://claude.com/claude-code)
<!-- devin-review-badge-begin -->

---

<a href="https://app.devin.ai/review/dyad-sh/dyad/pull/2821"
target="_blank">
  <picture>
<source media="(prefers-color-scheme: dark)"
srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img
src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1"
alt="Open with Devin">
  </picture>
</a>
<!-- devin-review-badge-end -->

---------
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>

#skip-bb
<!-- devin-review-badge-begin -->

---

<a href="https://app.devin.ai/review/dyad-sh/dyad/pull/2855"
target="_blank">
  <picture>
<source media="(prefers-color-scheme: dark)"
srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img
src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1"
alt="Open with Devin">
  </picture>
</a>
<!-- devin-review-badge-end -->

## Summary
- Updated Supabase prompt to emphasize using the `execute_sql` tool for
database operations
- Changed documentation to clarify that users should not manually create
migration files
- Added warning that migration files in `supabase/migrations/` are
automatically generated

## Test plan
- Review the updated prompt content in `src/prompts/supabase_prompt.ts`
- Verify that the guidance is clear about using execute_sql tool
- Ensure migration file warning is prominent

🤖 Generated with [Claude Code](https://claude.com/claude-code)
<!-- devin-review-badge-begin -->

---

<a href="https://app.devin.ai/review/dyad-sh/dyad/pull/2829"
target="_blank">
  <picture>
<source media="(prefers-color-scheme: dark)"
srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img
src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1"
alt="Open with Devin">
  </picture>
</a>
<!-- devin-review-badge-end -->

closes #2804 
<!-- devin-review-badge-begin -->

---

<a href="https://app.devin.ai/review/dyad-sh/dyad/pull/2841"
target="_blank">
  <picture>
<source media="(prefers-color-scheme: dark)"
srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img
src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1"
alt="Open with Devin">
  </picture>
</a>
<!-- devin-review-badge-end -->

---------
Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com>

Handle connection-drop and retry behavior in local agent IPC handling.
Add an end-to-end scenario that verifies recovery after a temporary
local agent disconnect.
Align test fixtures and snapshots for local-agent reconnection behavior.

## Summary
- Bump @ai-sdk/openai dependency in package.json from ^3.0.25 to
^3.0.36.
- Refresh package-lock.json to match the updated resolution.

## Test plan
- npm run fmt
- npm run lint:fix
- npm run ts

🤖 Generated with [Claude Code](https://claude.com/claude-code)
<!-- devin-review-badge-begin -->

---

<a href="https://app.devin.ai/review/dyad-sh/dyad/pull/2812"
target="_blank">
  <picture>
<source media="(prefers-color-scheme: dark)"
srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img
src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1"
alt="Open with Devin">
  </picture>
</a>
<!-- devin-review-badge-end -->

<!-- devin-review-badge-begin -->

---

<a href="https://app.devin.ai/review/dyad-sh/dyad/pull/2789"
target="_blank">
  <picture>
<source media="(prefers-color-scheme: dark)"
srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img
src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1"
alt="Open with Devin">
  </picture>
</a>
<!-- devin-review-badge-end -->

---------
Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: Will Chen <willchen90@gmail.com>

#skip-bb
<!-- devin-review-badge-begin -->

---

<a href="https://app.devin.ai/review/dyad-sh/dyad/pull/2811"
target="_blank">
  <picture>
<source media="(prefers-color-scheme: dark)"
srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img
src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1"
alt="Open with Devin">
  </picture>
</a>
<!-- devin-review-badge-end -->

## Summary
- Refactored the TokenBar's percentage and context window display to use
a cleaner flex-based layout
- Added a centered separator dot between percentage and context window
values
- Improved visual alignment with `items-center` and consistent spacing

## Test plan
- Open a chat to view the TokenBar component
- Verify the token percentage and context window display are properly
aligned
- Check that the separator dot appears with correct styling

🤖 Generated with [Claude Code](https://claude.com/claude-code)
<!-- devin-review-badge-begin -->

---

<a href="https://app.devin.ai/review/dyad-sh/dyad/pull/2783"
target="_blank">
  <picture>
<source media="(prefers-color-scheme: dark)"
srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img
src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1"
alt="Open with Devin">
  </picture>
</a>
<!-- devin-review-badge-end -->

---------
Co-authored-by: Will Chen <willchen90@gmail.com>
Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com>
Co-authored-by: claude[bot] <41898282+claude[bot]@users.noreply.github.com>

## Summary
- Add new skill definition for  under .
- Document required workflow for adding model configurations from
official docs.

## Test plan
- npm run fmt
- npm run lint:fix
- npm run ts

🤖 Generated with [Claude Code](https://claude.com/claude-code)
<!-- devin-review-badge-begin -->

---

<a href="https://app.devin.ai/review/dyad-sh/dyad/pull/2786"
target="_blank">
  <picture>
<source media="(prefers-color-scheme: dark)"
srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img
src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1"
alt="Open with Devin">
  </picture>
</a>
<!-- devin-review-badge-end -->

---------
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>

<!-- devin-review-badge-begin -->

---

<a href="https://app.devin.ai/review/dyad-sh/dyad/pull/2810"
target="_blank">
  <picture>
<source media="(prefers-color-scheme: dark)"
srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img
src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1"
alt="Open with Devin">
  </picture>
</a>
<!-- devin-review-badge-end -->

Fixes #2790 

## Summary
- Return SQL execution output from execute_sql tool instead of
discarding it
- Surface migration write failures alongside SQL execution results
- Preserve existing migration behavior while improving feedback for
callers

## Test plan
- npm run fmt
- npm run lint:fix
- npm run ts

🤖 Generated with [Claude Code](https://claude.com/claude-code)
<!-- devin-review-badge-begin -->

---

<a href="https://app.devin.ai/review/dyad-sh/dyad/pull/2797"
target="_blank">
  <picture>
<source media="(prefers-color-scheme: dark)"
srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img
src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1"
alt="Open with Devin">
  </picture>
</a>
<!-- devin-review-badge-end -->

## Summary
- **attach_image.spec.ts**: Fix flaky aria snapshot for
`upload-to-codebase` test by using regex patterns for non-deterministic
button text and removing race-dependent "extra files edited outside of
Dyad" assertion
- **context_compaction.spec.ts**: Replace brittle aria snapshots with
order-independent element assertions since compaction restructures
messages non-deterministically between runs

## Test plan
- [x] `attach_image.spec.ts` passes 30/30 runs with `--repeat-each=30`
- [x] `context_compaction.spec.ts` passes 20/20 runs with
`--repeat-each=10`
- [x] Full test suite (`npm test`) passes (842/842 tests)
- [x] Lint checks pass (`npm run fmt && npm run lint:fix && npm run ts`)

🤖 Generated with [Claude Code](https://claude.com/claude-code)
<!-- devin-review-badge-begin -->

---

<a href="https://app.devin.ai/review/dyad-sh/dyad/pull/2793"
target="_blank">
  <picture>
<source media="(prefers-color-scheme: dark)"
srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img
src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1"
alt="Open with Devin">
  </picture>
</a>
<!-- devin-review-badge-end -->

---------
Co-authored-by: claude[bot] <41898282+claude[bot]@users.noreply.github.com>
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>

## Summary
- cap `pr-review-responder` auto re-request attempts at 3 retries
- mark PRs with `cc:failed` and `needs-human:review-issue` when retry
limit is reached
- fix strict TypeScript errors in GitHub/local agent handlers so `npm
run ts` passes

## Test plan
- npm run fmt && npm run lint:fix && npm run ts
- npm test

🤖 Generated with [Claude Code](https://claude.com/claude-code)
<!-- devin-review-badge-begin -->

---

<a href="https://app.devin.ai/review/dyad-sh/dyad/pull/2798"
target="_blank">
  <picture>
<source media="(prefers-color-scheme: dark)"
srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img
src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1"
alt="Open with Devin">
  </picture>
</a>
<!-- devin-review-badge-end -->

---------
Co-authored-by: Will Chen <willchen90@gmail.com>

## Summary
- Add planning document for implementing a new `web_fetch` tool for
local agent mode
- This tool will fetch and read website content when users share URLs,
available to all users (free + Pro)
- Unlike the existing Pro-only `web_crawl` tool, `web_fetch` performs
direct local HTTP fetch at zero infrastructure cost

## Test plan
- Manual review of the plan document to ensure implementation strategy
is clear
- Implementation will follow the detailed testing strategy outlined in
the plan

🤖 Generated with [Claude Code](https://claude.com/claude-code)
<!-- devin-review-badge-begin -->

---

<a href="https://app.devin.ai/review/dyad-sh/dyad/pull/2801"
target="_blank">
  <picture>
<source media="(prefers-color-scheme: dark)"
srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img
src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1"
alt="Open with Devin">
  </picture>
</a>
<!-- devin-review-badge-end -->

<!-- devin-review-badge-begin -->

---

<a href="https://app.devin.ai/review/dyad-sh/dyad/pull/2803"
target="_blank">
  <picture>
<source media="(prefers-color-scheme: dark)"
srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img
src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1"
alt="Open with Devin">
  </picture>
</a>
<!-- devin-review-badge-end -->

## Summary
Add persistence layer for todos so they survive across multiple turns in
a chat session. Todos are now saved to disk after each update and
automatically loaded and injected into the conversation context when a
new turn begins.

## Key Changes
- **New module `todo_persistence.ts`**: Provides utilities to save,
load, and delete todo JSON files stored in `.dyad/todos/<chatId>.json`
  - `getTodosFilePath()`: Constructs the file path for a chat's todos
- `saveTodos()`: Persists todos to disk with metadata (updatedAt
timestamp)
- `loadTodos()`: Loads previously saved todos, gracefully handling
missing or corrupted files
  - `deleteTodos()`: Removes the todos file when all todos are completed

- **Updated `local_agent_handler.ts`**: 
  - Load persisted todos at the start of each turn
- Emit loaded todos to the renderer immediately so the UI reflects them
- Initialize the agent context with persisted todos instead of an empty
array
- Inject a synthetic system message reminding the LLM of incomplete
todos from the previous turn, following the same pattern as existing
todo reminder logic
  - Ensure `.dyad/` directory is gitignored (idempotent operation)

- **Updated `update_todos.ts`**:
  - Persist todos to disk after each tool execution
  - Delete the todos file when all todos are marked as completed

## Implementation Details
- Todos are stored as JSON with an `updatedAt` timestamp for potential
future use
- File I/O errors are logged as warnings but don't interrupt the agent
flow (graceful degradation)
- The synthetic message injection for persisted todos only occurs if
there are incomplete todos, avoiding unnecessary context pollution
- The `.dyad/` gitignore step is idempotent and aligns with existing
patterns in the codebase

<!-- devin-review-badge-begin -->

---

<a href="https://app.devin.ai/review/dyad-sh/dyad/pull/2713"
target="_blank">
  <picture>
<source media="(prefers-color-scheme: dark)"
srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img
src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1"
alt="Open with Devin">
  </picture>
</a>
<!-- devin-review-badge-end -->

---------
Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: Mohamed Aziz Mejri <mohamedazizmejri@Mohameds-Mac-mini.local>

## Summary
- Compute a stable Monaco `path` value from `appId` and `filePath` in
`FileEditor`
- Pass that `path` into the editor so file models are keyed consistently
across renders/apps

## Test plan
- npm run fmt
- npm run lint:fix
- npm run ts
- npm test

🤖 Generated with [Claude Code](https://claude.com/claude-code)
<!-- devin-review-badge-begin -->

---

<a href="https://app.devin.ai/review/dyad-sh/dyad/pull/2781"
target="_blank">
  <picture>
<source media="(prefers-color-scheme: dark)"
srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img
src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1"
alt="Open with Devin">
  </picture>
</a>
<!-- devin-review-badge-end -->

---------
Co-authored-by: Will Chen <willchen90@gmail.com>

closes #2576 
<!-- devin-review-badge-begin -->

---

<a href="https://app.devin.ai/review/dyad-sh/dyad/pull/2757"
target="_blank">
  <picture>
<source media="(prefers-color-scheme: dark)"
srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img
src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1"
alt="Open with Devin">
  </picture>
</a>
<!-- devin-review-badge-end -->

---------
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>

## Summary
- Refactor `useParseRouter` hook to improve code organization and
maintainability
- Add comprehensive unit tests for the `useParseRouter` hook covering
navigation state parsing, route matching, and edge cases

## Test plan
- Run `npm test` to verify all 842 tests pass
- The new tests in `src/__tests__/useParseRouter.test.ts` cover various
navigation scenarios

🤖 Generated with [Claude Code](https://claude.com/claude-code)
<!-- devin-review-badge-begin -->

---

<a href="https://app.devin.ai/review/dyad-sh/dyad/pull/2780"
target="_blank">
  <picture>
<source media="(prefers-color-scheme: dark)"
srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img
src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1"
alt="Open with Devin">
  </picture>
</a>
<!-- devin-review-badge-end -->
Co-authored-by: Will Chen <willchen90@gmail.com>
Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com>

## Summary
- Add step 8 to the `pr-push` skill that automatically adds the
`cc:request` label for non-trivial PRs to trigger code review
- This prevents issues like lint failures from slipping through when the
agent doesn't fully execute lint checks (e.g., due to token limits in
long CI runs)
- Renumbered subsequent steps (9-11) accordingly

## Test plan
- [ ] Verify the `pr-push` skill SKILL.md has the new step 8 with clear
trivial/non-trivial criteria
- [ ] Verify step numbering is correct (1-11)
- [ ] Run `pr-push` on a non-trivial PR and confirm `cc:request` label
is added

🤖 Generated with [Claude Code](https://claude.com/claude-code)
<!-- devin-review-badge-begin -->

---

<a href="https://app.devin.ai/review/dyad-sh/dyad/pull/2773"
target="_blank">
  <picture>
<source media="(prefers-color-scheme: dark)"
srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img
src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1"
alt="Open with Devin">
  </picture>
</a>
<!-- devin-review-badge-end -->
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>

Scope Import button locators to the GitHub URL tab panel to prevent
strict mode violations when both tab panels render Import buttons
simultaneously. Verified: 60/60 passes after fix.
<!-- devin-review-badge-begin -->

---

<a href="https://app.devin.ai/review/dyad-sh/dyad/pull/2775"
target="_blank">
  <picture>
<source media="(prefers-color-scheme: dark)"
srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img
src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1"
alt="Open with Devin">
  </picture>
</a>
<!-- devin-review-badge-end -->
Co-authored-by: claude[bot] <41898282+claude[bot]@users.noreply.github.com>
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>

<!-- This is an auto-generated description by cubic. -->
## Summary by cubic
Enable planning_questionnaire in normal agent mode and enforce a
clarify‑before‑plan workflow with explicit examples. Also adds a
one‑pass todo follow‑up so the agent completes remaining tasks before
stopping.

- **New Features**
  - planning_questionnaire is available in both plan and normal modes.
- Clarify step: call planning_questionnaire to ask 1–3 focused questions
for vague requests; skip when specific; STOP until the user replies.
Always use it for new app/major feature requests.
- One-pass todo follow-up: if a turn ends with incomplete todos, inject
a short reminder and run another pass to finish them.

- **Bug Fixes**
- Always stop after planning_questionnaire in both modes to prevent
repeated tool calls.
- Gate plan-only tools: write_plan and exit_plan are plan‑mode only;
questionnaire remains available in normal mode.

<sup>Written for commit c6a67ea7e6981e86418dddeaa245b918a22261e9.
Summary will update on new commits.</sup>

<!-- End of auto-generated description by cubic. -->


<!-- devin-review-badge-begin -->

---

<a href="https://app.devin.ai/review/dyad-sh/dyad/pull/2566"
target="_blank">
  <picture>
<source media="(prefers-color-scheme: dark)"
srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img
src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1"
alt="Open with Devin">
  </picture>
</a>
<!-- devin-review-badge-end -->

---------
Co-authored-by: Mohamed Aziz Mejri <mohamedazizmejri@Mohameds-Mac-mini.local>
Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com>

## Summary
- Hides the "Fix with AI" button in the `DyadOutput` component while a
response is being streamed
- Prevents users from triggering conflicting AI fix requests during an
ongoing stream
- Uses the existing `isStreamingByIdAtom` to track streaming state per
chat

## Test plan
- Open the app and trigger an error/warning in the output
- Start a streaming response
- Verify the "Fix with AI" button is hidden during streaming
- Verify the button reappears after streaming completes

🤖 Generated with [Claude Code](https://claude.com/claude-code)
<!-- devin-review-badge-begin -->

---

<a href="https://app.devin.ai/review/dyad-sh/dyad/pull/2771"
target="_blank">
  <picture>
<source media="(prefers-color-scheme: dark)"
srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img
src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1"
alt="Open with Devin">
  </picture>
</a>
<!-- devin-review-badge-end -->
Co-authored-by: Will Chen <willchen90@gmail.com>
Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com>

#skip-bb
<!-- devin-review-badge-begin -->

---

<a href="https://app.devin.ai/review/dyad-sh/dyad/pull/2770"
target="_blank">
  <picture>
<source media="(prefers-color-scheme: dark)"
srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img
src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1"
alt="Open with Devin">
  </picture>
</a>
<!-- devin-review-badge-end -->

## Summary
- Fixed flaky `refresh app` test in `refresh.spec.ts` by adding
pre-refresh snapshot verification and post-refresh content wait
- The test was failing ~20% of the time because the iframe body would
remain empty after clicking refresh
- Added `snapshotPreview()` before body removal to ensure initial
content is loaded
- Added `not.toHaveText("")` wait after refresh to ensure content
reloads before snapshotting

## Flaky test analysis

Scanned 1 main branch CI run (with available artifacts) + 9 PR comments
from wwwillchen/wwwillchen-bot. Top flaky spec files by occurrence:

| Spec File | Occurrences | Status |
|:---|:---:|:---|
| setup_flow.spec.ts | 13 | Already fixed (PR #2740) |
| debugging_logs.spec.ts | 6 | Stable (passed 50/50) |
| select_component.spec.ts | 6 | Stable (passed 60/60) |
| setup.spec.ts | 5 | Stable (passed 10/10) |
| switch_versions.spec.ts | 5 | Stable (passed 20/20) |
| **refresh.spec.ts** | **5** | **Fixed in this PR** |
| toggle_screen_sizes.spec.ts | 4 | Stable (passed 30/30) |
| undo.spec.ts | 3 | Stable (passed 30/30) |

## Verification
- `refresh.spec.ts` passed 40/40 runs with `--repeat-each=10` after the
fix (was 38/40 before)

## Test plan
- [x] Verify `refresh.spec.ts` passes consistently with
`--repeat-each=10`
- [x] No application code changes (test-only fix)

🤖 Generated with [Claude Code](https://claude.com/claude-code)
<!-- devin-review-badge-begin -->

---

<a href="https://app.devin.ai/review/dyad-sh/dyad/pull/2759"
target="_blank">
  <picture>
<source media="(prefers-color-scheme: dark)"
srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img
src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1"
alt="Open with Devin">
  </picture>
</a>
<!-- devin-review-badge-end -->

---------
Co-authored-by: claude[bot] <41898282+claude[bot]@users.noreply.github.com>
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>

## Summary
- Changed the Claude Deflake E2E workflow schedule from daily to
weekdays only (Monday-Friday at 10:00 UTC)
- This reduces unnecessary CI runs on weekends when less attention is
paid to flaky test results

## Test plan
- Verify the cron expression `0 10 * * 1-5` is valid and represents
weekdays only
- The workflow can still be triggered manually via `workflow_dispatch`

🤖 Generated with [Claude Code](https://claude.com/claude-code)
<!-- devin-review-badge-begin -->

---

<a href="https://app.devin.ai/review/dyad-sh/dyad/pull/2761"
target="_blank">
  <picture>
<source media="(prefers-color-scheme: dark)"
srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img
src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1"
alt="Open with Devin">
  </picture>
</a>
<!-- devin-review-badge-end -->
Co-authored-by: Will Chen <willchen90@gmail.com>
Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com>

## Summary

- **refresh.spec.ts**: Fix race condition where the "refresh app" test
would fail because the iframe body was empty after refresh. Added
explicit waits for the iframe body to have content both before removing
it and after clicking refresh, preventing the `toMatchAriaSnapshot` from
matching against an empty `<body></body>`.
- **themes_management.spec.ts**: Fix strict mode violations where
`getByLabel('Description (optional)')` resolved to 2 elements (one on
the Manual tab, one on the AI tab). Replaced ambiguous label selectors
with specific element ID selectors (`#manual-description`,
`#manual-prompt`, `#ai-description`). Also scoped edit dialog selectors
to the dialog role to avoid similar ambiguity.

## Verification

Both spec files were verified with `--repeat-each=10` (0 retries) after
fixes:
- `refresh.spec.ts`: 40/40 passed (was 34/40 before fix)
- `themes_management.spec.ts`: 50/50 passed (was 49/50 before fix)

## Flaky test scan results

Scanned 2 CI runs on main + 11 PRs by wwwillchen/wwwillchen-bot. Most
previously flaky tests (setup_flow, setup, select_component,
debugging_logs, switch_versions, etc.) appear to have been fixed in
prior commits. Only refresh.spec.ts and themes_management.spec.ts were
still reproducing locally.

## Test plan

- [x] `refresh.spec.ts` passes 10x repeat-each with 0 retries
- [x] `themes_management.spec.ts` passes 10x repeat-each with 0 retries
- [x] Formatting, linting, and type checks pass

🤖 Generated with [Claude Code](https://claude.com/claude-code)
<!-- devin-review-badge-begin -->

---

<a href="https://app.devin.ai/review/dyad-sh/dyad/pull/2752"
target="_blank">
  <picture>
<source media="(prefers-color-scheme: dark)"
srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img
src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1"
alt="Open with Devin">
  </picture>
</a>
<!-- devin-review-badge-end -->

---------
Co-authored-by: claude[bot] <41898282+claude[bot]@users.noreply.github.com>
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>

## Summary
- Adds comprehensive planning document for first-class NeonDB
integration at `plans/neondb-integration.md`
- Generated via swarm-to-plan session with PM, UX, and Engineering
agents who analyzed the idea, debated trade-offs, and incorporated user
decisions
- Covers full-stack design: agent tools, system prompt, UI changes,
security model, and example code for Neon + Drizzle ORM + NextAuth.js

## Test plan
- [ ] Review the plan document for completeness and accuracy
- [ ] Verify alignment with existing Supabase integration architecture
- [ ] Confirm example code patterns are correct

🤖 Generated with [Claude Code](https://claude.com/claude-code)
<!-- devin-review-badge-begin -->

---

<a href="https://app.devin.ai/review/dyad-sh/dyad/pull/2748"
target="_blank">
  <picture>
<source media="(prefers-color-scheme: dark)"
srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img
src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1"
alt="Open with Devin">
  </picture>
</a>
<!-- devin-review-badge-end -->
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>

#skip-bb
<!-- devin-review-badge-begin -->

---

<a href="https://app.devin.ai/review/dyad-sh/dyad/pull/2743"
target="_blank">
  <picture>
<source media="(prefers-color-scheme: dark)"
srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img
src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1"
alt="Open with Devin">
  </picture>
</a>
<!-- devin-review-badge-end -->

## Summary
- Adds a new Claude Code skill that monitors GitHub Actions workflow
runs from the past day
- Identifies severe or consistent failures and files GitHub issues for
actionable problems
- Includes detailed instructions for classifying failures, investigating
issues, and filing reports

## Test plan
- The skill can be manually triggered via `/dyad:check-workflows` or run
on a daily schedule at 13:00 UTC
- Check the generated GitHub issue (if any) to verify the classification
and reporting logic works correctly
- Verify that expected failures (like Nightly Runner Cleanup, cascading
CI failures) are properly filtered out

🤖 Generated with [Claude Code](https://claude.com/claude-code)
<!-- devin-review-badge-begin -->

---

<a href="https://app.devin.ai/review/dyad-sh/dyad/pull/2738"
target="_blank">
  <picture>
<source media="(prefers-color-scheme: dark)"
srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img
src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1"
alt="Open with Devin">
  </picture>
</a>
<!-- devin-review-badge-end -->
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>

## Summary
- Deflake `setup_flow.spec.ts` which was the most flaky test file (20
occurrences across 10 CI runs and 5 PRs)
- Wrap initial "Setup Dyad" heading check in `toPass()` with
`Timeout.MEDIUM` to handle slow page renders in CI
- Wrap `page.reload()` in `toPass()` retry to handle intermittent
`ERR_FILE_NOT_FOUND` errors in Electron
- Expand Node.js accordion section before checking for install button
(section not auto-expanded after reload with mock)
- Use `force: true` on "Continue" button click to avoid accordion
overlap pointer interception
- Replace hardcoded timeouts with `Timeout.MEDIUM` constants for
CI/local adaptability

## Test plan
- [x] Ran `setup_flow.spec.ts` with `--repeat-each=10` (30 tests total)
- all passed
- [x] Previously failed 6/10 times on "node.js install flow" test, now
passes 10/10
- [x] Other top flaky specs (context_manage, chat_input, edit_code,
logs_server, themes_management, fix_error) passed all repeats locally
without changes - their CI flakiness is environment-specific
- [x] All 819 unit tests pass
- [x] Lint and type checks pass

🤖 Generated with [Claude Code](https://claude.com/claude-code)
<!-- devin-review-badge-begin -->

---

<a href="https://app.devin.ai/review/dyad-sh/dyad/pull/2740"
target="_blank">
  <picture>
<source media="(prefers-color-scheme: dark)"
srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img
src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1"
alt="Open with Devin">
  </picture>
</a>
<!-- devin-review-badge-end -->
Co-authored-by: claude[bot] <41898282+claude[bot]@users.noreply.github.com>
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>

## Summary

Add a comprehensive planning document outlining an initiative to improve
the speed of landing PRs for maintainers by optimizing the CI pipeline,
automating the fix-and-retry loop, and improving developer feedback
during wait times.

The plan covers:
- CI pipeline optimizations for self-hosted macOS ARM64 runners
(caching, job consolidation)
- Auto-trigger fix loop for privileged author PRs (remove label
requirement)
- Tiered test selection based on changed files
- PR status watcher for terminal notifications
- Unified `/pr` skill entry point
- Move pr-review-responder to self-hosted runners

The plan is structured in four implementation phases: instrumentation,
CI quick wins, auto-fix loop enhancements, and developer experience
improvements. It includes risk assessments, testing strategies, and
decision logs.

## Test plan

- Verify the document is correctly formatted and renders as markdown
- Review the content for clarity and alignment with project goals
- Confirm all phases, user stories, and technical components are clearly
described

🤖 Generated with [Claude Code](https://claude.com/claude-code)
<!-- devin-review-badge-begin -->

---

<a href="https://app.devin.ai/review/dyad-sh/dyad/pull/2739"
target="_blank">
  <picture>
<source media="(prefers-color-scheme: dark)"
srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img
src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1"
alt="Open with Devin">
  </picture>
</a>
<!-- devin-review-badge-end -->

## Summary
- Add ADR 0001 for host capability interface decisions.
- Add ADR 0002 for cloud runtime topology and ADR 0003 for authz model.
- Add ADR index and architecture updates for cross-platform planning.
- Add desktop/mobile/web unification plan document in plans/.

## Test plan
- npm run fmt
- npm run lint:fix
- npm run ts

🤖 Generated with [Claude Code](https://claude.com/claude-code)

<!-- devin-review-badge-begin -->

---

<a href="https://app.devin.ai/review/dyad-sh/dyad/pull/2737"
target="_blank">
  <picture>
<source media="(prefers-color-scheme: dark)"
srcset="https://static.devin.ai/assets/gh-open-in-devin-review-dark.svg?v=1">
<img
src="https://static.devin.ai/assets/gh-open-in-devin-review-light.svg?v=1"
alt="Open with Devin">
  </picture>
</a>
<!-- devin-review-badge-end -->

---------
Co-authored-by: wwwillchen-bot <theodorewilsonchen@gmail.com>