Insights
Articles on how teams of specialists and agents agree on what to build, and factual updates from Loft.
-
Update
Loft 2.4: see whether your agent read the specification first
Yard now counts what your coding agents read and edit on each branch, and shows for each stage whether the specification was consulted before the code was first edited. It ships with open-source Loft 2.4, in the plugin for Claude Code and Codex.
-
Article
Tasks are derived from the change
Coding agents take a ticket as the unit of work. Loft takes a change as two states of your product's specification: as published, and as it should be. The tasks can always be worked out from the two states. The two states can never be worked back out of the tasks.
-
Update
Loft 2.3: Yard is available
Yard is where you and your coding agents meet while they work. It keeps a log on every branch of what they planned, asked, decided, and did, and a page on your machine shows all of it live. It ships with open-source Loft 2.3, in the plugin for Claude Code and Codex.
-
Article
Yard: the record of how a change was built
A specification says what the product must do, and the trace shows which tests check it. Yard keeps the third record on every branch: what your coding agent planned, asked, and decided, and what you answered. It ships with open-source Loft, in its plugin for Claude Code and Codex.
-
Update
Loft 2.2: A status pane, a term list, and test links that notice rewrites
Loft 2.2 shows where a specification stands while a coding agent works on it, lets a specification name its concepts in a term list, and reports a test rewritten after its link to an acceptance criterion was reviewed.
-
Article
Three months under a 53-hour release: building Loft with Claude Code
Open-source Loft went public 53 hours after its first commit. Under that sat three months of building Loft V1 with Claude Code around one idea: agents need a plant, a line of deterministic gates, more than a faster factory.
-
Update
Loft 2.1: Import OpenSpec projects and render them back
Loft 2.1 converts the main specs of an OpenSpec project into a Loft specification and writes a Loft specification back out as OpenSpec specs. A team using OpenSpec can start from the specs it already has and keep an OpenSpec copy while it adopts Loft.
-
Article
How good are OpenSpec specs?
We scored 87 public OpenSpec projects against OpenSpec's own writing rules and the classic tests of a good requirement, with two blind graders reading a sample. The format is followed almost everywhere. The writing is where projects differ, and it gets worse as specs grow.
-
Update
Loft 2.0: The specification format and command line are open source
Loft 2.0 releases the Loft Specification Format and the loft command line under the Apache 2.0 license. A person or a small team can keep a product contract in their own git repository, write it with a coding agent, and check it without a server.
-
Article
How to Check AI-Generated Software Against Acceptance Criteria
AI-generated code can look complete and still miss the intended behavior. Check it against the ratified requirements, criterion by criterion, and leave acceptance to an authorized person.
-
Article
How Loft Supports Product-Shaped Engineering
AI lets engineers work across more of the product lifecycle. Loft supports that broader ownership with a ratified product contract, explicit authority, and evidence for human acceptance.
-
Update
Loft 1.0.5 to 1.0.7: Agent-assisted specification review
Loft now gives teams and their agents a structured way to review a chosen product specification, surface potentially related assertions, and return findings to people for judgement.
-
Update
Loft 1.0.2 to 1.0.4: Production signals shape the product
Loft is being built through its own emerging Software Plant: production signals expose friction, people decide what should change, and the product specification and implementation evolve together.
-
Update
Loft 1.0.0 and 1.0.1: Multi-tenant Loft
Loft now supports separate customer workspaces with tenant-specific access controls, operator administration, and retrieval scoped to each workspace.
-
Article
When Every Role Gets Faster, Why Doesn't Software Delivery?
AI can make every specialist faster without making software delivery faster. A shared, human-ratified contract helps work compound across product, engineering, and assurance.