Conversation
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
commit: |
❌ Dogfood Test: failed |
|
✅ Expect tests passed Test output |
|
✅ Expect tests passed Test output |
|
✅ Expect tests passed Test output |
|
✅ Expect tests passed Test output |
|
✅ Expect tests passed Test output |
…planning and change-analysis guidance
…and update related functionality in prompts and flow storage
…amps for markers and improve state management
4092c5f to
a732368
Compare
Test Results✅ Website Test: passed10 passed, 0 failed out of 10 steps — 243s
Session Recording |
There was a problem hiding this comment.
1 issue found across 22 files
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/supervisor/src/flow-storage.ts">
<violation number="1" location="packages/supervisor/src/flow-storage.ts:36">
P2: `note` is derived from raw fields instead of resolved saved values, so it can duplicate `instruction` or `expectedOutcome` after fallback.</violation>
</file>
Reply with feedback, questions, or to request a fix. Tag @cubic-dev-ai to re-run a review.
packages/shared/src/models.ts
Outdated
| title: string; | ||
| instruction: string; | ||
| expectedOutcome: string; | ||
| routeHint?: string; |
There was a problem hiding this comment.
we shouldn't need SavedFlow and SavedFlowStep when we already have the ExecutedTestPlan?
There was a problem hiding this comment.
2 issues found across 4 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/shared/src/analytics/analytics.ts">
<violation number="1" location="packages/shared/src/analytics/analytics.ts:113">
P2: `machineId()` is now called on every analytics event instead of once at initialization, causing unnecessary repeated async work on the capture hot path.</violation>
</file>
<file name="packages/agent/src/build-session-meta.ts">
<violation number="1" location="packages/agent/src/build-session-meta.ts:12">
P2: Return `undefined` when no session metadata is present; the current implementation always returns `{}`, so `_meta` is always sent.</violation>
</file>
Reply with feedback, questions, or to request a fix. Tag @cubic-dev-ai to re-run a review.
… and system prompt
…y is only captured when enabled
There was a problem hiding this comment.
4 issues found across 13 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="apps/cli/src/utils/push-step-state.ts">
<violation number="1">
P2: Using `.find` here can return a stale `RunFinished` event; select the latest one so viewer status/summary reflect the final run state.</violation>
</file>
<file name="packages/shared/src/analytics/analytics.ts">
<violation number="1" location="packages/shared/src/analytics/analytics.ts:103">
P1: Computing `machineId()` at service initialization with `Effect.orDie` can fail Analytics service startup. Keep this non-fatal (or recover to a fallback ID) so telemetry issues don't break initialization.</violation>
</file>
<file name="packages/shared/src/models.ts">
<violation number="1">
P1: `tool_call_update` should finalize pending text markers before adding tool events. Otherwise step/run markers in trailing agent text can be permanently skipped once `ToolResult`/`ToolProgress` is appended.</violation>
</file>
<file name="packages/supervisor/src/reporter.ts">
<violation number="1">
P2: Use the last `RunFinished` event, not the first one, so report summaries reflect the final run state.</violation>
</file>
Reply with feedback, questions, or to request a fix. Tag @cubic-dev-ai to re-run a review.
Made-with: Cursor # Conflicts: # packages/shared/src/prompts.ts
… Actions - Updated buildSessionMeta function to include provider parameter and return specific metadata for claude when isGitHubActions is true. - Added tests to verify behavior for non-claude providers and handling of system prompts. - Removed unnecessary route hints and notes from prompts test for clarity.
No description provided.