launchpad

Collaborative brainstorming with codebase research, structured dialogue, and design document capture. NEVER writes code.

Meta-orchestrator for autonomous execution pipeline. Chains /lp-inf → /lp-review → /lp-resolve-todo-parallel → /lp-test-browser → /lp-ship → compound-learning.sh.

Stage changes, run quality gates, generate a conventional commit message, and optionally create a PR with CI monitoring

Dispatches copy review agents from review_copy_agents in agents.yml. Shell command — downstream projects populate the agent list.

Reads copy brief from section spec and provides copy context for design builds. Shell command — downstream projects extend with copy agents/skills.

Comprehensive 14-dimension design alignment audit against DESIGN_SYSTEM.md. Covers visual hierarchy, spacing, typography, color, alignment, components, iconography, motion, empty states, loading states, error states, dark mode, density, and accessibility. Does NOT modify code.

Captures live implementation state and compares against Figma design specs. Produces a structured review report with severity ratings. Does NOT modify code.

Iterative design improvement agent. Takes screenshots, analyzes design quality, implements ONE improvement per cycle, and repeats. Loads frontend-design, web-design-guidelines, and responsive-design skills for creative direction and compliance.

Responsive design audit: 6 checks covering breakpoints, mobile-first patterns, touch targets, overflow, container queries, and WCAG responsive requirements. Returns findings with P1/P2/P3 severity. Does NOT modify code.

Quick UI audit: 5 checks covering visual hierarchy, density, consistency, Jobs filter, and responsive behavior. Returns findings with P1/P2/P3 severity and file:line fix locations. Does NOT modify code.

Process skill for structured brainstorming sessions. Guides collaborative idea exploration through progressive questioning, approach comparison, and design document capture. Loaded by /lp-brainstorm and /lp-kickoff.

Stage changes, run quality gates (tests, linting, type checks, pre-commit hooks), generate a conventional commit message, and optionally create a PR with CI monitoring. Triggers on: commit changes, commit this, create commit, push and commit, commit my work, ready to commit.

Process skill for structured problem documentation. Defines 14-category taxonomy, YAML schema, and resolution templates for docs/solutions/. Loaded by /lp-learn.

Creates new Claude Code agents or converts existing skills into agents. Produces production-grade agent definitions with 8-section body structure, least-privilege tool assignment, and registration in CLAUDE.md/AGENTS.md. Triggers on: create agent, new agent, build agent, convert skill to agent, turn this into an agent.

Creates new Claude Code skills using the 7-phase Meta-Skill Forge methodology. Produces quality-validated skills with progressive disclosure, evaluation loops, and contrarian analysis. Use when creating a new skill, teaching Claude a new workflow, building a custom agent capability, or automating a repeatable task. Triggers on: create skill, new skill, build skill, teach Claude, make a skill for.