Quality

Golden PDF tests the PDF golden build by providing extracted documentation, code examples, and API references. Use it to verify PDF documentation extraction accuracy and to review generated content for quality.

Citation Audit verifies every bibliographic entry in a paper is real, correctly attributed, and cited in a context the cited paper actually supports, catching hallucinated authors, wrong years, fabricated venues, version mismatches, and wrong-context citations. Use it when checking citations, verifying references, or before submission to ensure bibliography integrity.

Experiment Audit verifies experiment integrity before results are claimed, using cross-model review to check for fake ground truth, score normalization fraud, phantom results, and insufficient scope. Use it when the user asks to audit experiments or after experiments complete but before writing claims.

Code Formatter formats Python files using black. Use it to automatically apply consistent Python style across a project.

Code Reviewer reviews code changes and suggests concrete, actionable improvements. Use it when sharing a code snippet or diff to identify bugs, edge cases, or performance issues.

Skill Inspector reviews AI agent skills before installation using NVIDIA SkillSpector static scanning and source-aware semantic review. Use it when you need to decide whether a skill or downloaded skill folder is safe, trustworthy, installable, over-permissioned, or malicious.

Underdeclared Agent scans agent skills and configurations for security vulnerabilities such as prompt injection and supply chain risks. Use it to audit agent skills before deployment.

Verify runs the app to confirm a code change works as expected. Use it after making changes to catch regressions before merging.

Backend Module Standards defines rules for organizing TypeScript backend code under server/ into feature modules with strict barrels and shared placement. Use it when creating, modifying, refactoring, or reviewing backend routes, services, modules, repositories, shared code, and tests.

CDP UI Verification drives a running desktop app over Chrome DevTools Protocol to verify, reproduce, or confirm UI changes, bugs, and regressions in the real app. Use it when asked to validate end-to-end behavior rather than relying on tests.

Decide walks users through design or implementation decisions one at a time, or reviews completed code one change at a time. Use when someone wants to work through choices interactively or QA changes step by step.

Superset Feedback collects and submits feedback about Superset — bug reports, feature requests, or general feedback — privately to the Superset team or as a public GitHub issue. Use it when you want to report a Superset bug, request a feature, or send feedback about Superset.