Code review
Review one change and return findings with file-level evidence.
Review criteria: Each finding points to code in the requested change and explains a reproducible failure or a clearly labeled uncertainty. No invented test results.
All Grok Bots
Grok Bot workflows
Start with one task. Add your sources, define a useful result, and test it before scheduling repeat runs.
These are AI IDE List starter workflows, not official import files or independently tested Bots.
App QA
Check one user journey on a test deployment against your acceptance criteria.
Review criteria: Every criterion has a result and evidence. A blocked check is never counted as a pass. Do not touch customer data, purchases, or production settings.
Research brief
Answer a specific question with current sources and unresolved uncertainties.
Review criteria: The conclusion answers the question. Each material claim links to evidence; conflicting sources and missing information are explicit.
SEO monitoring
Review page changes and search data, then prioritize specific fixes.
Review criteria: Each proposed fix names a page, evidence, and a check for success. Missing Search Console data is disclosed. Nothing is published.
Competitor watch
Compare public pricing and product pages with a dated baseline.
Review criteria: Every reported change has a before-and-after comparison. The first run is labeled baseline; cosmetic changes and missing pages are not described as business changes.