Prompt + runbook, ready to reuse

Mob Elaboration for Claude

The skill helps you choose a useful next step, prepare the intent and guide the session. It also recognises when a full workshop would add little. The full-session script is generated from the same source as the copyable prompt.

Version 0.3.0. An unofficial facilitation layer; rehearse before a live session.

Claude chat

Download the skill ZIP, then open Customize → Skills in Claude and upload it. Enable the skill before starting your conversation. Upload the ZIP as a skill, rather than attaching it to an ordinary message.

Download Claude skill

Start with: “Help me decide whether Mob Elaboration fits this work.” Or, when the team is ready: “Our team is here and we have an intent. Start Mob Elaboration.”

See Claude’s skill setup guide if Skills is unavailable in your account or workspace.

Claude Code

Add our marketplace, then install the plugin. Archive installation requires Claude Code 2.1.224 or later.

/plugin marketplace add https://mobelaboration.com/marketplace.json
/plugin install mob-elaboration@mobelaboration

Invoke /mob-elaboration:mob-elaboration, or ask Claude about Mob Elaboration. Follow any reload instruction shown by Claude Code.

This marketplace is maintained by mobelaboration.com. It is not an Anthropic official or community-catalog listing.

Download the Claude Code plugin ZIP · Inspect the marketplace manifest · Package checksums

Choose a mode

  • Full session: team questions, accepted units and open decisions; up to 145 minutes including optional risks.
  • Preparation: an intent brief and missing evidence, before the team meets.
  • Mandate audit: a separate 30-minute decision-authority diagnostic. Read the workflow.
  • Split session: live questions and later explicit review. Pending drafts are never counted as accepted.
  • Follow-up: a retrospective and post-implementation outcomes.

Small clear work, solo implementation or a technical feasibility question may need a simpler approach. Check whether Mob Elaboration fits.

What to expect

The facilitator owns the timer. The team owns decisions. The skill records evidence, assumptions and missing approvals, and can save agreed documents when repository access is available. It does not implement the work or submit feedback automatically.

Package checks do not prove facilitation quality. Effectiveness has not been established by comparative trials. Run a rehearsal in your chosen Claude environment and review the artifacts before using them.

Mob Elaboration and AI-DLC are AWS’s methodology and terms. This skill, its mandate audit and its split-session adaptation are independently maintained by mobelaboration.com.