Spin up a PR code review sub-agent. Ranks bugs critical/high/medium/low, eliminates false positives, writes consolidated report, applies KISS/DRY fixes (main logic only, skip tests), re-reviews after every change, loops until done. Use when the user asks to review a PR, triage bugs, or run a code review fix loop.
Perform comprehensive codebase analysis and generate reports (usage: /analyze [full|security|performance])
Review the current diff or selected files (usage: /review [--last-diff|--target <expr>|--file <path>|files...] [--style <style>])
Use for ast-grep: ast-grep run, sg scan, sg test, sg new, new rule, sgconfig.yml, inline-rules, stdin, json, optional chaining, rule catalog, meta variables, pattern objects, nthChild stopBy, range field, metadata url, caseInsensitive glob, severity off, include metadata, rule order, kind pattern, positive rule, kind esquery, debug-query, static analysis, tree-sitter parser, pattern yaml api, search rewrite lint analyze, textual structural, ast cst, named unnamed, kind field, ambiguous pattern, effective selector, meta variable detection, lazy multi, strictness smart, relaxed signature, string fix, fix config, expandEnd, replace substring, toCase separatedBy, rewriter, rewrite joinBy, find patch, barrel import, ruleDirs testConfigs, libraryPath languageSymbol, dynamic injected, custom language, TREE_SITTER_LIBDIR, language injection, styled components, language alias, languageGlobs, expandoChar, napi parse, python api, programmatic API.
Install VT Code skills into ~/.agents/skills from a curated list or a GitHub repo path. Use when a user asks to list installable skills, install a curated skill, or install a skill from another repo, including private repos.
Guide for creating effective skills. This skill should be used when users want to create a new skill (or update an existing skill) that extends VT Code's capabilities with specialized knowledge, workflows, or tool integrations.
| name | cmd-command |
| description | Run a terminal command (usage: /command <program> [args...]) |
| disable-model-invocation | true |
| metadata | {"slash_alias":"/command","usage":"/command <program> [args...]","category":"tools","backend":"traditional_skill"} |
Interpret the user input as the raw command string that follows /command.
