Mad Max — Design skill for Claude Code
Unattended backlog-drain mode — runs executor → verifier → code-reviewer loops on all ready tasks, logs stuck/human-needed, never prompts.
How to install Mad Max
Installs to ~/.claude/skills/kesteva-soloflow-mad-max/SKILL.md
mkdir -p ~/.claude/skills/kesteva-soloflow-mad-max && curl -fsSL https://raw.githubusercontent.com/kesteva/soloflow/HEAD/commands/mad-max.md -o ~/.claude/skills/kesteva-soloflow-mad-max/SKILL.md Restart Claude Code, or start a new session, for it to be picked up.
What Mad Max does
description: Unattended backlog-drain mode — runs executor → verifier → code-reviewer loops on all ready tasks, logs stuck/human-needed, never prompts. Visual verify off; parallel by default. argument-hint: [optional: TASK-NNN TASK-NNN ... or IDEA-NNN to filter] allowed-tools: [Read, Write, Edit, Glob, Grep, Bash, Agent, AskUserQuestion]
/soloflow:mad-max
Unattended variant of `/soloflow:sprint`. Drains every ready task in the backlog through the full per-task quality loop (executor
Alternatives in Design
- Website Builder Setup — Build professional animated websites with Claude Code — no coding needed 313 ★
- Apple Design Skill — Cross-platform UI/UX design reviewer based on Apple HIG 202 ★
- Light Mode Theming — Add a light mode to contrast the default dark mode, and build the design system so that 143 ★
Full documentation available on GitHub
View Source RepositoryRelated Skills
Loop Builder
Claude Code skill: design file-based agent loops (vision/specs/plan/prompts + bash engine) so you verify the o
Style Toggle
Switch the active terse mode: ponytail, caveman, or off — never both at once.
Iterate Centroids Score
Run the long-horizon centroid-score kernel iterate loop autonomously — design 4 variants, JIT-compile, benchma
Iterate Topk
Run the long-horizon top-k kernel iterate loop autonomously — design 4 variants, JIT-compile, benchmark, gener
Design Loop
Run the full design iteration loop (harness component 06): planner → generator → four machine gates → critic,
Claude E
PTY-backed claude -p-style executor for Claude Code, packaged as the claude-e npm CLI
Related Agents
Karr Foundation Worker
karr-foundation worker — multi-board discovery, overview, agent command resolution, drain loops, per-repo lock
Loop Operator
Safety net for autonomous agent loops. Detects stalls, reduces scope on repeated failures, escalates to user w
QA Correctness
Broken-output QA gate. Verifies text is present and not cut off, no black frames, no overlap, no stuck/duplica
