ac

Browser QA testing — ad-hoc tests, bug reproduction, plan verification. Uses Playwright CLI.

Smart commit — preflight checks, convention detection, atomic commits. Delegates to git-master when available.

Execute an approved plan — parallel background agents or sequential

Flutter QA testing — ad-hoc tests, bug reproduction, plan verification, integration test execution, visual regression. Uses flutter-skill MCP.

Idea refinement — Socratic interview, ambiguity scoring, adversarial challenge, task generation. Supports --bulk and --loop.

Generate or enhance project CLAUDE.md — auto-discovers codebase, interviews developer, preserves custom sections.

Auto-generate .claude/rules/ — path-scoped coding conventions from project analysis. Re-run after codebase changes to update rules while preserving manually added conventions.

Mobile QA testing — ad-hoc tests, bug reproduction, plan verification, flow execution. Uses Maestro MCP.

Structured planning — classify, research, analyze, interview, generate, review

Interactive coding style analyzer — scans existing projects, interviews the developer, generates a personalized my-coding skill

Generate global CLAUDE.md — interviews developer, detects skills/MCP, produces orchestration config for every CC session.

Interactive writing style analyzer — scans existing content, interviews developer, generates my-language skill

Browser test executor — runs pre-built test cases via Playwright CLI shell commands, captures evidence, returns structured verdicts. Spawned by /ac:browser-qa.

Devil's advocate for proposals, ideas, and architecture decisions. Use when stress-testing an approach before committing — during ideation (Phase 4) or pre-plan analysis (Complex plans).

Codebase search specialist. Use proactively for internal lookups — files, patterns, relationships, architecture. Returns file:line references. Fire for any question involving 2+ modules or unfamiliar code.

Feasibility evaluator — assesses codebase fit, effort, and dependencies before committing to an approach. Use proactively during ideation (Phase 4) or pre-plan analysis (Complex plans).

Flutter test executor — runs test cases via flutter-skill MCP tools on iOS/Android/Web/Desktop, captures evidence, returns structured verdicts. Spawned by /ac:flutter-qa.

External documentation specialist. Use proactively for library, API, framework, and CLI tool knowledge — even well-known ones. Finds official docs via context7 MCP with WebSearch fallback. Every claim cites a source URL.

LSP code intelligence verifier. Use after code changes to check diagnostics and confirm symbol structure.

Mobile test executor — runs test cases via Maestro MCP tools on iOS/Android emulators, captures evidence, returns structured verdicts. Spawned by /ac:maestro-qa.

Pre-planning gap detector and post-planning quality auditor. Use before plan generation (directives) and after (gap/slop detection).

2-stage code reviewer — spec compliance against plan criteria, then code quality. Use after implementation before marking complete.

Deep cross-layer code review for complex plans. Checks hidden coupling, signature changes affecting unmodified callers, and architectural compliance. Use after code-reviewer passes for complex plans.

Adversarial plan reviewer — bias toward REJECT. Deep reference verification, AI-slop detection, cross-task dependency analysis, tier challenge. OKAY or REJECT. Use for Complex plans (mandatory) or Standard (opt-in via --deep-review).

Plan reviewer — blockers-only, approval bias. Verifies references, executability, QA scenarios, tier sanity. OKAY or REJECT. Use after plan generation for Standard+ plans.

Post-execution plan compliance auditor. Verifies done-when criteria, forbidden patterns, and scope fidelity. Returns APPROVE or REJECT. Use after implementation to verify plan compliance.

Plan step executor. Executes a single plan step — code changes, server operations, or infrastructure tasks. Reads context, implements precisely, verifies results. Model overridden by orchestrator per step tier (quick→haiku, mid→sonnet, senior→opus).

Create Claude Code agents with optimal structure, tool control, and output contracts. Use when building subagents, advisory agents, or execution workers for Claude Code plugins.

Browser QA workflow patterns and Playwright CLI integration. Loaded by /ac:browser-qa command.

Write CC-optimal CLAUDE.md and CLAUDE.local.md files. Use when creating, reviewing, or improving any CLAUDE.md — global, project-level, or local. Covers section authoring, compression, dedup boundaries, and quality scoring.

Create Claude Code commands with phase-based structure and agent delegation. Use when building /plugin:command slash commands for Claude Code plugins.

Flutter QA workflow patterns and flutter-skill MCP tool routing for Claude Code agents. Reference for /ac:flutter-qa command.

Maestro QA workflow patterns and MCP tool integration. Loaded by /ac:maestro-qa command.

Write CC-optimal prompts for skills, agents, commands, and rules. Use when creating or improving any Claude Code extension component, or when writing prompts that need to align with how CC models process instructions.

Create path-scoped .claude/rules/ for Claude Code projects. Use when adding coding conventions, framework patterns, or domain-specific rules that auto-inject when matching files are touched.

Create Claude Code skills with progressive disclosure architecture. Use when building reusable workflow skills, domain-specific skills, or optimizing existing skills for Claude Code plugins.