lev-os

Use when inventorying conversation state into durable artifacts, routing shovel-ready items toward exec, or deep-processing a brain dump.

Use when verified work is ready for QA, acceptance, checkpointing, learning, recommendation, and handoff.

Use when a long brainstorm or brain dump needs deep capture, prior-art routing, and durable next states.

Use when running execution-ready task slices, dispatching verified `lev exec` flows, or reviewing implementation work in the Exec lane.

Use when guiding decision interviews that produce a design artifact in .lev/pm/designs before proposal or implementation planning.

Use when designing Lev Pentagon harnesses, eval suites, acceptance criteria, adversarial probes, or proof gates for agentic features before implementation or promotion.

Use when turning an aligned design or captured idea into task DNA and execution YAML through the Plan lane.

Use when creating skills, converting docs/repos/PDFs to skills, installing external skills, auditing skill security, or merging skills

Agentic UX design language and pattern library. Establishes CLI-as-Prompt paradigm where every CLI output is a prompt to the next agent. Includes Enriched Response Protocol, vernacular (Self-Injection, Teachable Touch, Bi-Directional), and implementation patterns for validation gates and trust calibration.

Automates browser interactions for web testing, form filling, screenshots, and data extraction. Use when the user needs to navigate websites, interact with web pages, fill forms, take screenshots, test web applications, or extract information from web pages.

Agent-human interaction protocol for AI agents. Use when agents need human approvals, selections, research direction, or async feedback. Triggers on "human in the loop", "approval", "agent interaction", "get human input", or "wait for response".

Think and reason like a software architect. Analyze systems through quality attributes, trade-off analysis, and architectural drivers. Use for system design, tech selection, ADRs, architecture reviews, C4 descriptions, fitness functions, and PR review. Triggers: "architect this", "design system", "architecture review", "ADR", "trade-off analysis", "quality attributes", "fitness functions", "C4 diagram", "system design", "review PR", "review this", "/review".

Automates browser interactions for web testing, form filling, screenshots, and data extraction. Use when the user needs to navigate websites, interact with web pages, fill forms, take screenshots, or extract information from web pages.

Create Mermaid diagrams (activity, deployment, sequence, architecture) from text descriptions or source code. Use when asked to "create a diagram", "generate mermaid", "document architecture", "code to diagram", "create design doc", or "convert code to diagram". Supports hierarchical on-demand guide loading, Unicode semantic symbols, and Python utilities for diagram extraction and image conversion.

Helps users discover and install agent skills when they ask questions like "how do I do X", "find a skill for X", "is there a skill that can...", or express interest in extending capabilities. This skill should be used when the user is looking for functionality that might exist as an installable skill.

Multi-agent orchestrator for Claude Code. Use when user mentions gastown, gas town, gt commands, bd commands, convoys, polecats, crew, rigs, slinging work, multi-agent coordination, beads, hooks, molecules, workflows, the witness, the mayor, the refinery, the deacon, dogs, escalation, or wants to run multiple AI agents on projects simultaneously. Handles installation, workspace setup, work tracking, agent lifecycle, crash recovery, and all gt/bd CLI operations.

Generate interactive playgrounds for AgentPing components

Standards-enforcing validation gates for git commits. Catches what linters miss.