Skip to content

[T618] Fix scoped static-web verification#284

Merged
ai21z merged 1 commit into
v0.9.0-beta-devfrom
T618
May 31, 2026
Merged

[T618] Fix scoped static-web verification#284
ai21z merged 1 commit into
v0.9.0-beta-devfrom
T618

Conversation

@ai21z

@ai21z ai21z commented May 31, 2026

Copy link
Copy Markdown
Owner

Summary

  • Add scoped static-web problem classification so single-target CSS/JS repair turns only block on target-relevant verifier findings.
  • Preserve strict full-site/static-web creation checks and target-affecting failures.
  • Surface unrelated static-web issues as contextual facts and avoid overclaiming full-site coherence for scoped passes.

Verification

  • RED observed first for scoped CSS verification and outcome rendering before implementation.
  • ./gradlew.bat test --tests "dev.talos.runtime.verification.StaticTaskVerifierTest.scopedCssRewriteDoesNotFailOnUnrelatedMissingJavaScriptLink" --tests "dev.talos.runtime.verification.StaticTaskVerifierTest.scopedCssRewriteStillFailsWhenCssTargetIsEmpty" --tests "dev.talos.runtime.verification.StaticTaskVerifierTest.scopedCssRewriteStillFailsWhenHtmlDoesNotLinkCssTarget" --tests "dev.talos.runtime.verification.StaticTaskVerifierTest.scopedJavaScriptRewriteStillFailsWhenHtmlDoesNotLinkJavaScriptTarget" --tests "dev.talos.runtime.verification.StaticTaskVerifierTest.fullStaticWebCreateStillFailsWhenHtmlDoesNotLinkJavaScriptTarget" --tests "dev.talos.cli.modes.ExecutionOutcomeTest.postApplyScopedCssVerificationDoesNotOverclaimFullWebCoherence" --no-daemon
  • ./gradlew.bat test --tests "dev.talos.runtime.verification.StaticTaskVerifierTest" --tests "dev.talos.cli.modes.ExecutionOutcomeTest" --no-daemon
  • ./gradlew.bat test --tests "dev.talos.runtime.toolcall.RolefulIntentRecoveryRegressionTest" --no-daemon
  • ./gradlew.bat test --tests "dev.talos.runtime.outcome.EvidenceContainmentAnswerGuardTest" --tests "dev.talos.cli.modes.ReadEvidenceHandoffTest" --no-daemon
  • git diff --check
  • ./gradlew.bat validateArchitectureBoundaries --no-daemon
  • ./gradlew.bat check --no-daemon

@chatgpt-codex-connector chatgpt-codex-connector Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

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

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 9bb70f8fab

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

List<String> formProblems = StaticWebStructureVerifier.calculatorFormProblems(
contract.originalUserRequest(), selectors.html());
problems.addAll(formProblems);
staticWebProblems.addAll(formProblems);

Copy link
Copy Markdown

Choose a reason for hiding this comment

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

P1 Badge Keep calculator form failures blocking for scoped JS fixes

When the turn is scoped to a single JavaScript target, these calculator/form structure failures are now passed through StaticWebProblemScope. That scope only treats JavaScript-named problems as blocking, so messages such as Calculator/form task is missing a form or input container become contextual facts and the verifier can pass a request like “fix scripts.js so the BMI calculator works” even though the HTML lacks required inputs/buttons/results. These form checks were previously added directly to problems, so this weakens functional-web verification for scoped JS repairs.

Useful? React with 👍 / 👎.

@ai21z ai21z merged commit 308cdca into v0.9.0-beta-dev May 31, 2026
1 check passed
@ai21z ai21z deleted the T618 branch May 31, 2026 16:51
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.

1 participant