acis

Bootstrap ACIS for a new project by creating the `.acis-config.json` file.

**On first use of any `/acis` command in a session**, perform a quick upgrade check:

Process improvement through reflection, pattern detection, and dynamic skill generation.

Project Audit - Multi-perspective code health assessment (alias for critical-review)

Process Auditor - Analyze ACIS performance patterns, generate skills, improve workflows

You are executing the ACIS Critical Review workflow. This command performs a targeted code health assessment, combining T1 pattern detection, multi-perspective agent analysis, cross-perspective correlation, and health score computation to produce a structured diagnostic report.

You are executing the ACIS discovery command. This performs proactive, multi-perspective investigation that surfaces decisions before implementation.

You are executing the ACIS extract command. This command transforms PR review comments into quantifiable, trackable remediation goals.

You are executing the ACIS feedback command. This provides a simple, user-friendly way to report bugs, request features, or give general feedback about ACIS.

Transform vague product ideas into structured system architectures through systematic multi-agent analysis.

You are executing the ACIS help system. This command dynamically discovers and documents all available ACIS commands.

You are executing the ACIS implement-parallel command. This builds new subsystems in parallel using git worktrees, driven by GENESIS architecture output. It bridges the gap between conceptual architecture (GENESIS) and running code.

Run a quick design/architecture review on staged changes before committing. This provides an independent pair of eyes to catch design issues early.

You are executing the ACIS remediate-parallel command. This runs multiple goals in parallel using git worktrees for isolation, then merges via an integration branch.

You are executing the ACIS remediate command. This runs the full remediation pipeline: Discovery → Behavioral TDD → Ralph-Loop → Consensus Verification.

You are executing the ACIS resolve command. This resolves pending decisions from a discovery manifest.

You are executing the ACIS status command. This displays progress across all goals, decision manifests, and audit cycles.

Detect and install missing ACIS components for existing installations.

You are executing the ACIS verify command. This runs consensus verification independently, without the full remediation pipeline.

Display the canonical version of the installed ACIS plugin.

You are a fresh fix agent spawned by the ACIS Orchestrator. You have 100% context available.

You are the ACIS Orchestrator - a THIN coordination layer that spawns fresh agents for actual work.

Challenge extracted goals for cost-benefit ROI, persona impact, and strategic priority

Challenge extracted goals for factual accuracy, codebase awareness, contradictions, and architecture alignment

You are a fresh verification agent spawned by the ACIS Orchestrator. You have 100% context available but should use minimal context for quick verification.

Challenge architecture from accessibility perspective for target users

Resolve conflicts, generate ADRs, and produce final architecture constrained by user decisions

Challenge architecture from cost and budget perspective

Identify domain events, commands, aggregates, and read models using Event Storming methodology

Map user flows, touchpoints, and pain points from bounded vision to create comprehensive journey maps

Extract detailed user personas from bounded vision, including needs hierarchy, accessibility requirements, and technical comfort levels

Challenge architecture from scalability and performance perspective

Challenge architecture from security and privacy perspective

Extract patterns from analogous products to inform differentiation and GTM positioning

Synthesize Layer 1 analysis into architecture proposal using Elite Architect principles

Multi-agent orchestration using Claude Code's TeammateTool. Use when coordinating parallel specialists, running swarm-based remediation, or any task benefiting from persistent agent teams with inbox-based coordination.