Closed
Conversation
📝 WalkthroughWalkthroughReplaces static Changes
Sequence DiagramsequenceDiagram
participant User
participant SettingsModal
participant SettingsStore
participant getEffectiveAutoRunPrompt
participant BatchRunnerModal
User->>SettingsModal: Toggle/set auto-run prompt override
SettingsModal->>SettingsStore: setAutoRunDefaultPromptOverride(customPrompt)
SettingsStore->>SettingsStore: Persist to window.maestro.settings
Note over User,BatchRunnerModal: When BatchRunnerModal initializes
BatchRunnerModal->>getEffectiveAutoRunPrompt: getEffectiveAutoRunPrompt()
getEffectiveAutoRunPrompt->>SettingsStore: Read autoRunDefaultPromptOverride
alt Override set (non-empty)
getEffectiveAutoRunPrompt-->>BatchRunnerModal: Return trimmed override
else
getEffectiveAutoRunPrompt-->>BatchRunnerModal: Return built-in default
end
BatchRunnerModal->>BatchRunnerModal: Initialize/reset prompt to effective value
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~22 minutes Possibly related PRs
Suggested labels
Suggested reviewers
Poem
🚥 Pre-merge checks | ✅ 1 | ❌ 2❌ Failed checks (1 warning, 1 inconclusive)
✅ Passed checks (1 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
Add a new Auto Run tab in Settings with a toggle switch to customize the default prompt used for all Auto Run / Playbook executions. When off, the built-in default prompt is shown read-only. When on, the prompt becomes editable and the override applies globally to any execution that does not have a per-playbook custom prompt. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
scriptease
force-pushed
the
autorun-settings
branch
from
09807ab to
1f43704
Compare
Greptile SummaryThis PR adds a global Auto Run default prompt override in Settings, allowing users to replace the built-in Confidence Score: 5/5Safe to merge — feature is well-contained, all execution paths use the centralised All remaining findings are P2 (whitespace-only edge case in the toggle UI). The core feature — persisting a global override and applying it across wizard, symphony, and playbook flows — is implemented correctly. Tests are updated and the validation warning from the prior review thread is now present. No files require special attention.
|
| Filename | Overview |
|---|---|
| src/renderer/components/Settings/SettingsModal.tsx | Adds new "Auto Run" tab with toggle + textarea for editing the global prompt override; includes task-reference validation warning. |
| src/renderer/hooks/batch/batchUtils.ts | Adds getEffectiveAutoRunPrompt() (reads override with trim/fallback) and validateAgentPromptHasTaskReference() — both are clean and correct. |
| src/renderer/stores/settingsStore.ts | Adds autoRunDefaultPromptOverride field and its setter; pattern is consistent with other settings in the store. |
| src/renderer/hooks/batch/usePlaybookManagement.ts | Uses getEffectiveAutoRunPrompt() as fallback when loading a playbook with no saved prompt — correct behavior. |
| src/renderer/hooks/symphony/useSymphonyContribution.ts | Switched from hardcoded default to getEffectiveAutoRunPrompt() for batch run config prompt. |
| src/renderer/hooks/wizard/useWizardHandlers.ts | Switched from hardcoded default to getEffectiveAutoRunPrompt() for wizard-launched batch run config. |
| src/tests/renderer/hooks/useSymphonyContribution.test.ts | Mocks getEffectiveAutoRunPrompt to decouple test from settings store — correct test isolation. |
| src/tests/renderer/hooks/useWizardHandlers.test.ts | Mocks getEffectiveAutoRunPrompt for test isolation — correct approach. |
| src/renderer/hooks/settings/useSettings.ts | Adds autoRunDefaultPromptOverride / setAutoRunDefaultPromptOverride to the UseSettingsReturn interface — clean addition. |
| src/renderer/components/BatchRunnerModal.tsx | Uses getEffectiveAutoRunPrompt() as the initial prompt default when no per-session prompt exists — correct integration. |
| src/renderer/hooks/batch/index.ts | Exports new utilities (getEffectiveAutoRunPrompt, validateAgentPromptHasTaskReference) from the batch module. |
Flowchart
%%{init: {'theme': 'neutral'}}%%
flowchart TD
A([User opens Settings → Auto Run tab]) --> B{Override toggle on?}
B -- "No (autoRunDefaultPromptOverride = '')" --> C[Read-only textarea\nshows built-in default]
B -- "Yes (non-empty string)" --> D[Editable textarea\nshows override value]
D --> E{validateAgentPromptHasTaskReference?}
E -- false --> F[Warning shown:\n'prompt does not reference\nMarkdown tasks']
E -- true --> G[No warning]
H([Batch run triggered]) --> I[getEffectiveAutoRunPrompt]
I --> J{autoRunDefaultPromptOverride\n.trim non-empty?}
J -- Yes --> K[Use override prompt]
J -- No --> L[Use built-in autorunDefaultPrompt]
K --> M([BatchRunnerModal / Wizard / Symphony / Playbook])
L --> M
subgraph "Per-session priority"
N{activeSession.batchRunnerPrompt\nnon-empty?} -- Yes --> O[Use per-session prompt]
N -- No --> I
end
Reviews (2): Last reviewed commit: "fix: add task reference validation warni..." | Re-trigger Greptile
![greptile-apps[bot]](https://avatars.githubusercontent.com/in/867647?s=60&v=4){kind=small}
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 6
🧹 Nitpick comments (2)
scripts/showcase/seed/data/maestro-groups.json (1)
5-5: Consider using Title Case for group names.The group names are in ALL CAPS (
"BOOKMARKS","PROJECTS"), which may not reflect typical user usage patterns. For more realistic seed data, consider Title Case (e.g.,"Bookmarks","Projects") unless the ALL CAPS styling is intentional for the showcase.📝 Suggested naming refinement
{ "id": "00000000-0000-4000-a000-000000000001", - "name": "BOOKMARKS", + "name": "Bookmarks", "emoji": "⭐", "collapsed": false }, { "id": "00000000-0000-4000-a000-000000000002", - "name": "PROJECTS", + "name": "Projects", "emoji": "📁", "collapsed": false }Also applies to: 11-11
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@scripts/showcase/seed/data/maestro-groups.json` at line 5, The group names in the seed JSON are written in ALL CAPS (e.g., "BOOKMARKS", "PROJECTS"); update these values to Title Case (e.g., "Bookmarks", "Projects") in the JSON array so seeded data reads more realistic for users—locate the string values for the group name keys (the "name" fields containing "BOOKMARKS" and "PROJECTS") and replace them with their Title Case equivalents consistently across the file.src/renderer/components/BatchRunnerModal.tsx (1)
156-164: Potential stale state after global setting change.
effectiveDefaultis computed once when the component renders. If the user modifies the global Auto Run prompt in Settings while this modal is open, then returns and clicks "Reset", the prompt would reset to the new default (line 357 callsgetEffectiveAutoRunPrompt()fresh), butisModifiedandhasUnsavedChanges(lines 409-410) would still compare against the oldeffectiveDefaultcaptured at mount.This is a minor edge case (requires both modals open), but could cause confusing UI state where Reset appears to work but badges/buttons don't update correctly.
♻️ Optional: Use a getter function for consistent behavior
- const effectiveDefault = getEffectiveAutoRunPrompt(); - const [prompt, setPrompt] = useState(initialPrompt || effectiveDefault); + const [prompt, setPrompt] = useState(initialPrompt || getEffectiveAutoRunPrompt()); const [variablesExpanded, setVariablesExpanded] = useState(false); const [savedPrompt, setSavedPrompt] = useState(initialPrompt || ''); const [promptComposerOpen, setPromptComposerOpen] = useState(false); const textareaRef = useRef<HTMLTextAreaElement>(null); // Track initial prompt for dirty checking - const initialPromptRef = useRef(initialPrompt || effectiveDefault); + const initialPromptRef = useRef(initialPrompt || getEffectiveAutoRunPrompt());Then update comparisons to call the function:
- const isModified = prompt !== effectiveDefault; - const hasUnsavedChanges = prompt !== savedPrompt && prompt !== effectiveDefault; + const effectiveDefault = getEffectiveAutoRunPrompt(); + const isModified = prompt !== effectiveDefault; + const hasUnsavedChanges = prompt !== savedPrompt && prompt !== effectiveDefault;Moving
effectiveDefaultcloser to usage ensures fresh values for comparisons while keeping initialization stable.🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@src/renderer/components/BatchRunnerModal.tsx` around lines 156 - 164, The component captures effectiveDefault once at mount which can become stale if the global Auto Run prompt changes while the modal is open; update checks and initial state to use a fresh value by replacing the cached constant with a getter that calls getEffectiveAutoRunPrompt() when needed (or call getEffectiveAutoRunPrompt() inline in comparisons), and ensure initialPromptRef is initialized consistently (e.g., set initialPromptRef.current = initialPrompt || getEffectiveAutoRunPrompt() at mount) so isModified and hasUnsavedChanges (and the Reset handling) always compare against the current effective default; update references to effectiveDefault in functions that compute isModified/hasUnsavedChanges/reset logic to call the getter instead of using the stale effectiveDefault variable.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@scripts/showcase/generate-seed.js`:
- Around line 1016-1048: The settingsData object currently hardcodes
defaultShell: 'bash' which breaks Windows showcases; remove the defaultShell
property from settingsData in generate-seed.js (or, alternatively, call the
platform-aware getDefaultShell() from src/main/stores/defaults.ts if you prefer
to derive it at seed-time) so the seeded settings match the checked-in
maestro-settings.json and platform defaults; after changing the settingsData
object, regenerate the seed file so the showcase uses the platform-safe default
shell.
In `@scripts/showcase/launch.js`:
- Around line 25-29: The spawn call hardcodes MAESTRO_DEMO_DIR to
'/tmp/maestro-showcase', which breaks on Windows and duplicates the path in
scripts/showcase/setup.js; change the env value in the spawn invocation to a
platform-neutral directory built with require('os').tmpdir() and path.join
(e.g., path.join(os.tmpdir(), 'maestro-showcase')), and refactor so both
launch.js (the spawn env) and scripts/showcase/setup.js derive that same path
from a shared helper or the same computation (MAESTRO_DEMO_DIR constant) to
avoid drift; update the spawn invocation’s env to use that computed value
instead of the hardcoded string.
In `@scripts/showcase/plan.md`:
- Around line 148-153: Update the fenced code blocks that currently start with
``` to include a language tag (use text or markdown) so markdownlint stops
flagging them; specifically add a language to the fence containing the snippet
with "scripts/showcase/seed/data/" and the following "/tmp/maestro-showcase/"
block, and also apply the same change to the other fenced block between lines
235–259 that shows the showcase file layout. Make sure each opening
triple-backtick becomes ```text or ```markdown for those blocks.
In `@scripts/showcase/setup.js`:
- Around line 55-63: The current replaceInFile function performs raw string
replacements which corrupt JSON on Windows (unescaped backslashes like in a $CWD
path); change the logic to detect JSON files or attempt JSON.parse: if the file
is JSON or parses successfully, load the object, update only the specific fields
that contain the search token (traverse keys or check property values), then
write back using JSON.stringify(obj, null, 2) to preserve valid escaping; if
it's not JSON, fall back to the existing string replaceAll behavior. Apply the
same fix for the similar replacement logic referenced around the other block
(lines 97-110) so JSON values are always written safely.
In `@src/renderer/hooks/batch/batchUtils.ts`:
- Around line 16-19: In getEffectiveAutoRunPrompt, the override value from
useSettingsStore.getState().autoRunDefaultPromptOverride should treat
whitespace-only strings as empty; change the logic to trim the override (e.g.,
const override =
useSettingsStore.getState().autoRunDefaultPromptOverride?.trim()) and then
return override if non-empty, otherwise return autorunDefaultPrompt so
whitespace-only overrides no longer produce a blank prompt.
In `@src/renderer/hooks/groupChat/useGroupChatHandlers.ts`:
- Around line 320-323: The current logic treats 'agent-working' as moderatorFree
and can dispatch a new moderator turn while previous turn participants are still
running, which can overwrite the single pendingParticipantResponses set in the
main process; update the drain/dispatch checks so only groupChatState === 'idle'
is considered moderator-free (remove 'agent-working' from moderatorFree) in the
block using groupChatExecutionQueue and in the analogous block around the other
dispatch (the code handling lines ~630-646), ensuring new turns are not started
until the main process is made turn-aware and the previous
pendingParticipantResponses set is settled.
---
Nitpick comments:
In `@scripts/showcase/seed/data/maestro-groups.json`:
- Line 5: The group names in the seed JSON are written in ALL CAPS (e.g.,
"BOOKMARKS", "PROJECTS"); update these values to Title Case (e.g., "Bookmarks",
"Projects") in the JSON array so seeded data reads more realistic for
users—locate the string values for the group name keys (the "name" fields
containing "BOOKMARKS" and "PROJECTS") and replace them with their Title Case
equivalents consistently across the file.
In `@src/renderer/components/BatchRunnerModal.tsx`:
- Around line 156-164: The component captures effectiveDefault once at mount
which can become stale if the global Auto Run prompt changes while the modal is
open; update checks and initial state to use a fresh value by replacing the
cached constant with a getter that calls getEffectiveAutoRunPrompt() when needed
(or call getEffectiveAutoRunPrompt() inline in comparisons), and ensure
initialPromptRef is initialized consistently (e.g., set initialPromptRef.current
= initialPrompt || getEffectiveAutoRunPrompt() at mount) so isModified and
hasUnsavedChanges (and the Reset handling) always compare against the current
effective default; update references to effectiveDefault in functions that
compute isModified/hasUnsavedChanges/reset logic to call the getter instead of
using the stale effectiveDefault variable.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
Run ID: 0d2c4343-3c4c-4960-8d03-23f9b70bbeca
⛔ Files ignored due to path filters (1)
scripts/showcase/screenshot.pngis excluded by!**/*.png
📒 Files selected for processing (28)
CONTRIBUTING.mdTHEMES.mdpackage.jsonscripts/showcase/generate-seed.jsscripts/showcase/launch.jsscripts/showcase/plan.mdscripts/showcase/seed/data/group-chats/00000000-0000-4000-a000-00000000005b/metadata.jsonscripts/showcase/seed/data/maestro-agent-session-origins.jsonscripts/showcase/seed/data/maestro-claude-session-origins.jsonscripts/showcase/seed/data/maestro-groups.jsonscripts/showcase/seed/data/maestro-sessions.jsonscripts/showcase/seed/data/maestro-settings.jsonscripts/showcase/seed/data/maestro-window-state.jsonscripts/showcase/setup.jssrc/__tests__/renderer/hooks/useSymphonyContribution.test.tssrc/__tests__/renderer/hooks/useWizardHandlers.test.tssrc/main/group-chat/group-chat-router.tssrc/prompts/group-chat-participant-request.mdsrc/renderer/components/BatchRunnerModal.tsxsrc/renderer/components/Settings/SettingsModal.tsxsrc/renderer/hooks/batch/batchUtils.tssrc/renderer/hooks/batch/index.tssrc/renderer/hooks/batch/usePlaybookManagement.tssrc/renderer/hooks/groupChat/useGroupChatHandlers.tssrc/renderer/hooks/settings/useSettings.tssrc/renderer/hooks/symphony/useSymphonyContribution.tssrc/renderer/hooks/wizard/useWizardHandlers.tssrc/renderer/stores/settingsStore.ts
|
Tip: Greploop — Automatically fix all review issues by running Use the Greptile plugin for Claude Code to query reviews, search comments, and manage custom context directly from your terminal. |
There was a problem hiding this comment.
Actionable comments posted: 1
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@src/renderer/components/Settings/SettingsModal.tsx`:
- Around line 695-699: The toggle logic uses raw string truthiness for
autoRunDefaultPromptOverride (and the other similar override states) but runtime
uses trimmed values, so replace truthy checks with trimmed checks (e.g., use
autoRunDefaultPromptOverride?.trim() to decide enabled/disabled) and when
toggling set the state to either '' or autorunDefaultPrompt.trim() (or the
corresponding default trimmed value) via setAutoRunDefaultPromptOverride so
whitespace-only strings are treated as empty; apply the same change to the other
override state setters and checks in this component to ensure consistent
behavior.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
Run ID: b2998f84-1013-42bf-8f18-4bca25c61608
📒 Files selected for processing (11)
src/__tests__/renderer/hooks/useSymphonyContribution.test.tssrc/__tests__/renderer/hooks/useWizardHandlers.test.tssrc/renderer/components/BatchRunnerModal.tsxsrc/renderer/components/Settings/SettingsModal.tsxsrc/renderer/hooks/batch/batchUtils.tssrc/renderer/hooks/batch/index.tssrc/renderer/hooks/batch/usePlaybookManagement.tssrc/renderer/hooks/settings/useSettings.tssrc/renderer/hooks/symphony/useSymphonyContribution.tssrc/renderer/hooks/wizard/useWizardHandlers.tssrc/renderer/stores/settingsStore.ts
✅ Files skipped from review due to trivial changes (1)
- src/renderer/hooks/symphony/useSymphonyContribution.ts
🚧 Files skipped from review as they are similar to previous changes (5)
- src/tests/renderer/hooks/useWizardHandlers.test.ts
- src/tests/renderer/hooks/useSymphonyContribution.test.ts
- src/renderer/hooks/batch/usePlaybookManagement.ts
- src/renderer/hooks/batch/batchUtils.ts
- src/renderer/components/BatchRunnerModal.tsx
…verrides Show a warning in the Auto Run settings tab when the custom prompt does not reference Markdown tasks, matching the existing BatchRunnerModal validation. Also trim the override value so whitespace-only strings fall back to the built-in default instead of producing a blank prompt. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
scriptease
force-pushed
the
autorun-settings
branch
from
6d7bc41 to
8e3ec13
Compare
There was a problem hiding this comment.
🧹 Nitpick comments (1)
src/renderer/components/Settings/SettingsModal.tsx (1)
211-252: Consolidate tab order into one source of truthLine 212-252 duplicates tab ordering that is already represented in
TAB_ITEMS(Line 385-403). Keeping both lists in sync is easy to miss on future tab additions/removals.Refactor sketch
+ const TAB_ITEMS: Array<...> = [ ... ]; + const tabOrder = TAB_ITEMS.map((tab) => tab.id); useEffect(() => { if (!isOpen) return; const handleTabNavigation = (e: KeyboardEvent) => { - const tabs: Array<...> = FEATURE_FLAGS.LLM_SETTINGS ? [...] : [...]; - const currentIndex = tabs.indexOf(activeTab); + const currentIndex = tabOrder.indexOf(activeTab); if ((e.metaKey || e.ctrlKey) && e.shiftKey && e.key === '[') { e.preventDefault(); - const prevIndex = currentIndex === 0 ? tabs.length - 1 : currentIndex - 1; - setActiveTab(tabs[prevIndex]); + const prevIndex = currentIndex === 0 ? tabOrder.length - 1 : currentIndex - 1; + setActiveTab(tabOrder[prevIndex]); } else if ((e.metaKey || e.ctrlKey) && e.shiftKey && e.key === ']') { e.preventDefault(); - const nextIndex = (currentIndex + 1) % tabs.length; - setActiveTab(tabs[nextIndex]); + const nextIndex = (currentIndex + 1) % tabOrder.length; + setActiveTab(tabOrder[nextIndex]); } };Also applies to: 385-403
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@src/renderer/components/Settings/SettingsModal.tsx` around lines 211 - 252, The duplicated tab list in handleTabNavigation should be replaced with a single source of truth derived from TAB_ITEMS: compute the tabs array by mapping TAB_ITEMS to their tab keys/ids (e.g., TAB_ITEMS.map(item => item.key) or item.id depending on TAB_ITEMS shape) and then conditionally filter out 'llm' when FEATURE_FLAGS.LLM_SETTINGS is false; update handleTabNavigation to use that derived tabs array so order is maintained in one place and future tab changes only require edits to TAB_ITEMS.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Nitpick comments:
In `@src/renderer/components/Settings/SettingsModal.tsx`:
- Around line 211-252: The duplicated tab list in handleTabNavigation should be
replaced with a single source of truth derived from TAB_ITEMS: compute the tabs
array by mapping TAB_ITEMS to their tab keys/ids (e.g., TAB_ITEMS.map(item =>
item.key) or item.id depending on TAB_ITEMS shape) and then conditionally filter
out 'llm' when FEATURE_FLAGS.LLM_SETTINGS is false; update handleTabNavigation
to use that derived tabs array so order is maintained in one place and future
tab changes only require edits to TAB_ITEMS.
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
Run ID: e4d1ea2a-fe09-4daa-8083-8ba92d7f97ba
📒 Files selected for processing (2)
src/renderer/components/Settings/SettingsModal.tsxsrc/renderer/hooks/batch/batchUtils.ts
🚧 Files skipped from review as they are similar to previous changes (1)
- src/renderer/hooks/batch/batchUtils.ts
Collaborator
|
@scriptease this is neat; however, we have a pending PR that will make all of the prompts in Maestro editable, so that would supersede this one. |
Collaborator
|
Side note: if you change the auto-run prompt, it does persist for that agent. still changing the default is valuable of course. |
Author
|
This will return as part of: #197 |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Overwrite the default autorun prompt for all future agents and agents that don't have a customized prompt already.
My use case it removing the default push to github which takes 10 minutes for some projects, I am looking at you Maestro ;-)
Summary by CodeRabbit
New Features
Improvements