Interop 2027 is open for proposals, and the window closes on September 23rd. Developers who want a say in which web features browser vendors prioritize can use a coding agent already running in their terminal or IDE to identify gaps in their own projects and map them to the proposal list, instead of manually reading through GitHub issues.

Prompting an agent to surface feature gaps

The starting point is this prompt, which has the agent inventory web feature usage in the current project:

Audit this codebase for Interop 2027 alignment:

1.  **Scan the Project**:
   -   Run a Lighthouse "Baseline Features" audit using DevTools MCP.
   -   Scan files for polyfills, wrapper libs, `@supports` blocks, and custom JS workarounds.

2.  **Match open proposals**:
    -   Fetch open proposals using `gh issue list -R web-platform-tests/interop -l focus-area-proposal -s open`.
    -   List proposals that directly match features used or would be using if not for compat issues.

3.  **Find Unproposed Gaps**:
    -   Identify platform features we need that lack an Interop 2027 proposal.
    -   Verify each candidate has a stable spec (W3C/WHATWG), ≥1 engine implementation, WPT tests, and no vendor blockers.
    -   Refer to https://github.com/web-platform-tests/interop/blob/main/proposal_guide.md for more information on what makes a good proposal.

**Output**: A summary table of **Proposals to Upvote** (with estimated code/libs eliminated) and **Vetted Features to Propose** (with spec/WPT links). Sort both tables by highest impact in descending order, with impact measured as estimated code/libs eliminated.

Running it produces an analysis of what the project uses, matched against the list of open Interop proposals. The result falls into three buckets.

Features already in use

The agent reports features the codebase relies on, identified by web feature ID — async-clipboard is one example — along with each feature's Baseline status, where it appears in the code, and any polyfills or other gating mechanisms it detects around that usage. This is the context that makes the rest of the output actionable for a specific project.

Proposals to upvote

The most useful outcome is a list of features that matter to the codebase and already have open Interop proposals. Those can be upvoted directly.

Features to propose

Features relevant to the project that lack an existing proposal can be submitted by the developer. Two documents come first: what makes a good project proposal and the proposal guide in the Interop repository. If the feature clears the criteria, file an issue and select Focus area proposal. Note that the proposal submission rules require proposals to be written by humans.

Why it matters

Legacy code that developers would rather replace keeps accumulating while more features reach Baseline — and gaps that remain are opportunities for the Interop process to close. Collecting evidence from an actual project makes it easier to say which features and fixes would help daily workflows, and the agent handles the discovery step.

Submit or upvote before September 23rd.