Conversation
update syntax
Mesa DescriptionDescription
Testing
Description generated by Mesa. Update settings |
There was a problem hiding this comment.
Performed full review of 89aa20b...0d15518
Analysis
-
Non-Deterministic Template Ordering: The
GetSupportedTemplatesForLanguagefunction uses unordered map iteration, causing templates to appear in random order each time users run the command, creating an inconsistent UX. Templates should be sorted before display. -
Insufficient Validation: No verification that template directories exist in the embedded filesystem, required files are present, or that template metadata matches actual directories. This creates risk of runtime failures when templates are used.
-
Eliminated Test Coverage: Removal of
TestTemplatesleaves the template registry without automated validation, increasing risk of misconfiguration with the expanded template set. -
Overly Broad Embed Directive: Changing from
//go:embed all:typescriptto//go:embed all:*now embeds ALL files, including potentially sensitive or unnecessary hidden files, which could lead to bloated binaries or security issues. -
Suboptimal Error Handling: Error handling in
GetTemplateKeyFromValuelacks wrapped sentinel errors, making it difficult for callers to distinguish between different error types.
Tip
Help
Slash Commands:
/review- Request a full code review/review latest- Review only changes since the last review/describe- Generate PR description. This will update the PR body or issue comment depending on your configuration/help- Get help with Mesa commands and configuration options
102 files reviewed | 0 comments | Edit Agent Settings • Read Docs
update comment
0d15518 to
ed1d016
Compare
| await page.evaluate( | ||
| """(width, height) => {window.resizeTo(width, height)}""", | ||
| **self.browser_profile.window_size, | ||
| ) |
There was a problem hiding this comment.
Bug: Incorrect unpacking of arguments in page.evaluate call
The page.evaluate call uses **self.browser_profile.window_size which unpacks the dictionary as keyword arguments (width=1024, height=786). However, Playwright's page.evaluate method signature is evaluate(expression, arg=None) - it accepts a single positional arg parameter, not keyword arguments. This will raise a TypeError because evaluate() doesn't accept width and height as keyword parameters. The dictionary needs to be passed directly as a single argument, and the JavaScript expression updated to destructure it.
Description
onkernel/create-kernel-app/templatesTesting
Note
Adds comprehensive Python and TypeScript app templates (Computer Use, CUA, Stagehand, Gemini CUA, Magnitude, samples) and updates template embedding to include all assets.
computer-use(Anthropic),cua(OpenAI CUA),stagehand,gemini-cua(Gemini 2.5 + Stagehand),magnitude, and a basic sample app with Playwright integration.package.json,tsconfig.json), lockfiles, and utility code (tools, loops, utils).computer-use(Anthropic) with a full sampling loop, Playwright-based computer tools, and Kernel app entrypoint.cua(OpenAI) with Playwright computers (local + Kernel), agent loop, and utilities.pyproject.toml,uv.lock, READMEs, and _gitignore files.pkg/templates/templates.goto embedall:*so non-TS assets are included.Written by Cursor Bugbot for commit 97f7d38. This will update automatically on new commits. Configure here.