Skip to main content

Lamina vs spec tools

Spec Kit, Kiro, and similar tools structure engineering tasks and acceptance criteria. Lamina specs product-level states, roles, and flows, then feeds GraphVersion projection into your workflow.

Curated by Aryan Iyappan · Updated 2026-07-27

Who this comparison is for

You are in the right place if

  • Your spec has API tasks but permission matrices are missing.
  • Engineering specs pass but product behavior still breaks in the UI.
  • Spec Kit generated tasks from a one-line feature request with no product design.
  • Kiro workflow is green but guest and partner flows were never distinguished.

Context

Spec-driven development improved how agents break work into PR-sized chunks. The engineering spec layer assumes product behavior is already understood. In AI-assisted builds, that assumption fails often: tasks are well-formed while flows are wrong. Lamina sits upstream of engineering specs, not instead of them.

Common situations

Concrete cases from AI-shipping workflows. Each shows what the alternative handles and where validated graph knowledge and verification evidence help.

Spec Kit tasks for booking API

Situation
Spec Kit produced CRUD tasks and OpenAPI shapes. UI let guests pay while hold was expired because no task covered transition rules.
What Spec tools does
Breaks engineering work into implementable units with acceptance criteria.
What Lamina adds
Product-level state machine and pay preconditions before tasks are generated.
Takeaway
Engineering tasks need a behavior source document.

Kiro structured sprint on auth refactor

Situation
Backend auth migration, no new UI flows or role changes.
What Spec tools does
Kiro workflow coordinates implementation.
What Lamina adds
Skip Lamina unless UI permissions change.
Takeaway
Backend-only: spec tools alone may suffice.

OpenSpec enforces component tokens

Situation
Agentic spec tooling keeps components on design tokens. Partner submit gate still missing.
What Spec tools does
Component and token enforcement during build.
What Lamina adds
Domain permissions and verify on live partner flow.
Takeaway
Token enforcement is not product behavior design.

Comparison

Lamina vs spec tools (Spec Kit, Kiro)
DimensionSpec toolsLamina
Primary jobEngineering task structure and SDLCProduct behavior design
Layer in stackEngineering spec / implementation layerTransactional product graph + agent workflow
Primary outputTasks, APIs, acceptance criteriaValidated GraphVersion plus query projections
Writes app source?Agent implements from engineering specNo
Behavior spec (states, flows, permissions)Task-level criteria and API contractsState machines, actor permissions
Durable in-repo artifactSpec files in repoValidated GraphVersion plus query projections
Post-build verificationTests against engineering criteriaIndependent Persona Missions with runtime evidence
Repo and stack fitSpec-driven reposCLI plus portable skills in any repo and stack
Best whenBreaking work into engineering tasksDefining product rules before tasks
When you don't need LaminaBackend-only changeNo UI behavior to spec
Compose with Lamina?Implement from engineering spec/lamina-design → feed GraphVersion projection to Spec Kit

Spec tool roster

Lamina vs spec tools: when to use each
SituationSpec toolsLamina
Spec KitEngineering specs and agent tasksProduct behavior first
KiroStructured dev workflowresolved graph knowledge upstream
agentic-spec, nibComponent/token enforcementProduct-level flows
OpenSpec-style workflowsRepo spec conventionsGraphVersion as behavior input

Frequently asked questions

Which comes first: Lamina or Spec Kit?

Product behavior with /lamina-design first, then engineering tasks from GraphVersion projection.

Is GraphVersion projection duplicate of my engineering spec?

GraphVersion projection is product behavior handoff, not file-level tasks. Spec tools translate it into engineering work.

Spec tools generate tests. Does that replace verify?

Tests assert scenarios you defined. /lamina-verify discovers gaps before tests exist.

Comparisons map stack layers, durable outputs, and provenance-marked legacy examples. Updated when major tools change positioning.

Install Lamina with your agent.

Copy the repository-matched prompt. Your agent installs the standalone CLI, adds every Lamina skill for your active agent, configures supported passive rules, checks the setup, and tells you when to start a fresh session with /lamina-init.

Skills work across coding agents. Passive setup currently supports Codex, Claude Code, and Cursor.

  • Cursor
  • Claude Code
  • Codex
  • Gemini
  • Pi
  • and more
Install manually on macOS/Linux4 shell commands

Run these in your project, then start a fresh agent session. On Windows, use the PowerShell installer in the copied agent prompt.

curl -fsSL https://github.com/aryaniyaps/lamina/releases/latest/download/install.sh | sh
npx skills add aryaniyaps/lamina --skill '*' -a <active-agent> -y
lamina setup --agent <codex|claude-code|cursor>
lamina doctor --json