heya

The diagnostic workflow

This is Heya's specialty: take a WordPress or WooCommerce bug from a report to a proven, paste-ready answer. It runs on a disposable WordPress Playground or a development site you point it at. It never touches production, and it never posts anything. You post.

The shape

Heya runs four stages, each gated on evidence:

  1. Reproduce. Turn the report into a structured spec, build a clean environment, and confirm the bug actually happens, code-level first, then the browser. No evidence, no verdict.
  2. Diagnose. Classify the issue (a code bug, a config issue, a plugin or theme conflict, an environment or version problem, and so on), run the conflict test when it fits, and localize the likely cause. A fresh, skeptical verifier drops any claim it cannot ground.
  3. Remediate and verify. Propose the right kind of fix (a setting, a snippet, a patch, a version bump), ground it against the installed source, apply it in the disposable environment, and prove it with a dual oracle: the original reproduction now passes and a regression check still passes.
  4. Deliver. Aggregate everything into a triage report and a paste-ready comment carrying the decision bar, and write them to repro/<slug>/.

Triage a single issue

Give Heya the report, a ticket reference, or a log:

heya "triage WOO-1234: variation coupons apply to the parent price at checkout, WP 6.5, WC 8.7"

Heya intakes the issue (it can fetch a Linear or GitHub ticket through an MCP server, gh, or the web), runs the stages, and writes:

The decision bar

The comment is the product. Someone should be able to set a priority or close the ticket from it without re-investigating. It carries:

Triage a backlog

Point Heya at a list of issues or a view. It ranks them from their text alone, builds nothing yet, and writes a pick-list to pick-list.md, with one route per issue: ready-to-fix, triage-first, needs-info, or skip. When you confirm which to validate, it runs those through the single-issue flow, a few at a time.

The hard rules