Review the changes since a fixed point (commit, branch, tag, or merge-base) along two axes — Standards (does the code follow this repo's documented coding standards?) and Intent (does the code match the user prompt, Plannator plan, issue, task, PRD, commit message, or other source of intended behavior?). Runs both reviews in parallel sub-agents when the review is large enough and reports them side by side. Use when the user wants to review a branch, a PR, work-in-progress changes, or asks to "review since X".
Shared vocabulary for designing deep modules. Use when the user wants to design or improve a module's interface, find deepening opportunities, decide where a seam goes, make code more testable or AI-navigable, or when another skill needs the deep-module vocabulary.
Diagnosis loop for hard bugs and performance regressions. Use when the user says "diagnose"/"debug this", or reports something broken/throwing/failing/slow.
A relentless interview to sharpen a plan or design.
Use the installed wb CLI for persistent browser automation with compact page JSON and screenshots.
Structural refactoring and safe code rewrites with ast-grep across supported languages. Use for codemods, removing or replacing AST nodes (imports, docstrings, arguments), pattern-based migrations, and multi-file transformations when plain text search/replace is likely to be unsafe.
Selective hunk-level staging and commits with git-hunks. Use when the user asks to split changes into multiple logical commits, stage individual hunks instead of whole files, make surgical commits from a dirty worktree, or review changes at hunk granularity before committing. Also use when the user asks to "commit in hunks," "commit by hunk," or requests fine-grained staging.
Query Sumo Logic operational logs from the command line using sumocli. Use when the user asks to search Sumo logs, find production errors, count events by field, investigate a service/pod/host, or otherwise read data from Sumo Logic. Covers query syntax basics, time ranges, result modes (messages vs aggregation records), and output formats suitable for jq/awk/duckdb.