Documentation checked
Where Mob Elaboration fits among spec-driven tools
Spec-driven tools help turn intent into reviewable development artifacts. Mob Elaboration focuses on the conversation in which people clarify that intent. The approaches overlap and can work together.
This compares documented capabilities, not measured performance. The suggested uses below are our assessment, not a benchmark or a claim that other tools cannot support teamwork.
The category, briefly
Spec Kit and OpenSpec add structure around a coding agent. Kiro offers an integrated development environment across several surfaces. BMAD organises work through specialised roles and workflows. Tessl supplies reusable agent context and skills. AWS AI-DLC spans a wider delivery process. Follow the linked official documentation for each product’s current scope.
Our runbook is a particular implementation of an AI-guided team session. It uses a prompt in your existing chat; it does not build or execute software itself.
Side by side
On a phone, scroll the table sideways. Product names link to official sources. Suggested uses are editorial judgments.
| Approach / maintainer | What it provides | A possible fit | Cost | Keep in mind |
|---|---|---|---|---|
| GitHub Spec KitGitHub | A toolkit for specifying, planning, breaking down tasks and implementing with a supported coding agent. | Keeping requirements, plans and tasks as separate, reviewable artifacts. | Open-source toolkit; agent usage may cost extra. | Not limited to one spec file or one developer. Teams still own the requirements and review. |
| KiroAWS | An AI development environment with specs and steering across IDE, CLI and web surfaces. | An integrated path from requirements through design and tasks. | Free tier and paid plans; see current pricing. Pricing. | Adopting the IDE is not mandatory. Check feature availability on the surface you choose. |
| OpenSpecFission AI and contributors | A change-based workflow with proposals, spec deltas, designs and tasks. | Describing incremental changes to an existing system without rewriting its whole specification. | Open-source framework; agent usage may cost extra. | It also supports new projects. Specifications still need deliberate review and maintenance. |
| BMAD-METHODBMad community | Role-based agents and workflows for planning and building software. | Structuring product, architecture and implementation work with specialised AI assistance. | Open-source method; model or agent usage may cost extra. | AI roles do not confer business authority or replace stakeholder approval. Not every task needs the full workflow. |
| TesslTessl | An agent enablement platform with a registry, reusable skills and evaluation capabilities. | Sharing and evaluating the instructions and context supplied to coding agents. | Usage and plans vary; see current pricing. Pricing. | Its spec-driven methodology tile is one part of the offering. Describing all of Tessl as a spec-to-code compiler is misleading. |
| AWS AI-DLC workflowsAWS / AWS Labs | Structured software-delivery workflows that run through multiple coding-agent environments. | Connecting requirements, decisions, implementation and review across a delivery lifecycle. | Open-source workflow; model and infrastructure costs are separate. | Includes human approval gates. It is not accurately described as assuming a solo developer. |
| This Mob Elaboration runbookMobElaboration, unofficial | A facilitation runbook and session-runner prompt used in your own AI chat. | Helping a cross-functional team discuss intent, record unresolved decisions and review units of work. | Free prompt and runbook; your AI service and team time are separate. | Requires people with relevant knowledge and mandate. The 2 h 25 min timebox is this runbook’s design, not a guaranteed outcome. |
The distinction is emphasis, not exclusivity
Development tools can ask clarifying questions, record decisions and require human approval. A facilitated session does not have exclusive rights to those practices. Nor does any tool automatically guarantee that specifications and code remain consistent.
This runbook puts particular emphasis on who can make a decision: unresolved questions get an assumption, a named owner and a deadline. The architect or tech lead reviews unit boundaries before a unit is labelled autonomous. These are instructions to follow and verify, not technical enforcement.
A useful specification needs informed decisions. When authority or knowledge is missing, make that gap explicit; neither a prompt nor a meeting can supply it automatically.
A practical way to combine them
- Session. Use the session runner prompt to capture intent, answers, acceptance criteria and open decisions with the relevant people.
- Handoff. Save the approved artifacts in the repository. Ask your chosen tool to incorporate them into its own specification and task structure. This is a proposed manual handoff, not a built-in integration; review the conversion for omissions.
- Implementation. Start only work whose prerequisites are satisfied. Keep approvals, testing and code review in place, including for units labelled autonomous.
If you already have clear requirements and the authority to decide, a separate group session may add little value. If several people hold different parts of the answer, the session may be useful before or during a tool-based workflow.
Related techniques
Mob programming and ensemble programming provide the shared-working pattern behind the elaboration session. The connection is broader than coding: it includes discussion, design and learning together.
- Example Mapping explores a story through rules, examples and questions. Its question cards are related to an open-decisions board, but do not inherently include owners and deadlines.
- BDD and Gherkin can express behaviour through concrete Given / When / Then scenarios. That format alone does not make an acceptance criterion correct.
- EARS provides patterns for writing requirements. It helps structure sentences; it cannot resolve missing domain knowledge.
- Impact Mapping connects goals, actors, impacts and deliverables when the intended outcome needs clarification.
- EventStorming supports collaborative exploration of a business domain and can inform discussions of boundaries.
What this comparison can establish
Official documentation supports the product descriptions; it does not establish which approach produces better outcomes for your team. We do not present comparative trial results, verified adoption counts or a public session dataset here.
Tessl’s spec-driven development tile illustrates why categories overlap: requirements gathering and approval can also be packaged as reusable agent instructions. Evaluate the specific workflow you intend to use, rather than treating a product label as a guarantee.
Sources are linked beside the relevant descriptions and were checked on the date above. Features and pricing can change. This site is independent of the vendors listed and uses no affiliate links in this comparison.