Repository navigation
Conversation
yummybomb
force-pushed
the
hypeship/fandango-showtimes
branch
from
September 30, 2026 02:12
44cd460 to
fc0e532
Compare
yummybomb
marked this pull request as ready for review
September 30, 2026 14:52
yummybomb
added a commit
that referenced
this pull request
Oct 8, 2026
bench/ measures, per user intent on a site, how much more often and how much faster the same agent succeeds when the domain's read-only xwebmcp tools are registered in its Kernel browser. Each trial runs pi against one Kernel browser; both arms get the same model, prompt, kernel MCP server, tool allowlist and browser settings, and treatment also registers the domain's read-only tools at commits pinned for the run. run.mjs plans and freezes a run, gates it on a smoke test, runs treatment-only adoption trials and then balanced AB/BA pairs over parallel slots, and resumes by run id. grade.mjs grades answers by shape and rules; report.mjs prints per-intent screening labels with Wilson and Newcombe intervals, a paired speedup and sign test, and per-PR and per-run rollups. Includes intents for the domains on main and for #55, #58 and #69.
yummybomb
force-pushed
the
hypeship/fandango-showtimes
branch
from
October 29, 2026 15:55
07a175f to
c6ee118
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Safety boundary
All eight tools run in the current www.fandango.com page without navigation. No tool opens checkout, selects/reserves seats, validates selection, creates holds, changes cart/account state or purchases. Preparation is not applied or saved. Prices are nonfinal preview estimates; taxes/discounts/checkout parity are unverified. Unsupported identity shapes, seat-block members, admissions or surcharges fail closed. General-admission prices and consequential execution are absent: the runtime has no trusted approval path bound to an exact purchase, and a model-provided confirmation flag is not approval.
Validation
2de90754d451a42025914e158aefff83f91d4d02; subsequent commit only records README/registry metadata, leaving all reviewed tool files unchanged.Limitations
US public movie-ticket product only. First theater page for movie showtimes; catalog next_page is only a candidate. No seats are held or guaranteed; final charges, adjacency/orphan-seat rules and eligibility need manual rechecking. Cookie values changed during one independent preview probe with unproven attribution; no blanket storage-stability or hidden hold/cart guarantee. A separate headless invocation timed out with unknown outcome and was neither counted as success nor automatically retried. Real challenges, non-US/signed-in behavior, midnight cases, surcharges and consequential operations were not live-established. See the domain README for precise evidence and boundaries.
Review fixes
Status after these fixes: get_movie and list_movies stay
verified(unchanged source); the other six tools aredraftuntil fresh independent QA. The QA evidence above was for the earlier source.seat_offset/next_seat_offsetpage the whole map.offset/next_offset: a live NYC day listed 41 theaters on one site page, above the 40-theater limit.prepare_checkoutis renamedprepare_ticket_plan; it never opens checkout.lib/fandango-screening.jsinstead of copying it.canceled.