Skip to content

docs: draft agent-oriented linting paper#67

Draft
danielchen0 wants to merge 18 commits into
mainfrom
paper/agent-oriented-linting
Draft

docs: draft agent-oriented linting paper#67
danielchen0 wants to merge 18 commits into
mainfrom
paper/agent-oriented-linting

Conversation

@danielchen0
Copy link
Copy Markdown
Collaborator

Summary

Drafts an arXiv-style paper for laint around the idea of agent-oriented linting for generated JSX/TSX applications. The draft frames laint as a feedback-loop tool for catching framework-specific generated-app failures before build, preview, or runtime.

This is intentionally a draft PR: the paper has the system description, rule taxonomy, and evaluation plan, but still needs measured results before it should be treated as submission-ready.

Verification

  • make in paper/ builds main.pdf cleanly
  • Checked paper/main.log for undefined refs/citations and overfull/warning/error lines

Next

  • Add real authors/affiliations
  • Run the evaluation over generated app traces
  • Replace the evaluation-plan section with measured results
  • Add any venue-specific formatting if we target a workshop as well as arXiv

Comment thread paper/main.tex
\maketitle

\begin{abstract}
Large language model agents increasingly generate complete application code for web, mobile, and backend targets, but generated applications fail in ways that traditional linters do not emphasize. In particular, generated React, Next.js, Expo, React Native, and serverless TypeScript projects often fail through framework-specific integration mistakes: browser APIs used during server rendering, missing declarations that a component must run on the client, invalid routing conventions, mobile layout constraints, unsafe serverless loops, or platform-incompatible imports. These defects may be syntactically valid, type-correct, and invisible until build, preview deployment, device testing, or runtime.
Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

could include something about performance or smaller iteration cycles due to the feedback loop being tied to the agent's action lifecycle, not runtime/build time

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I know it's mentioned later on but might be nice to have as a hook idk if that's a thing in academic papers

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants