About software-factory
Software Factory is a skill for carrying out traceable software delivery work that connects product intent, technical design, and implementation. It guides Work Order execution and requires project-specific rules and referenced Blueprints to be read through the Software Factory MCP before authoring records or coding.
Key features
- 01Connects requirements, design, and delivery
- 02Guides Work Order execution
- 03Routes tasks to execution guides
- 04Requires record-specific writing rules
- 05Tracks checklist completion and skip reasons
Use cases
- 01Executing a repository Work Order
- 02Planning implementation during execution
- 03Creating or updating Factory records
- 04Reviewing work against execution procedures
MCP support
The skill uses the Software Factory MCP to retrieve project writing rules and referenced Blueprints. It also uses the MCP to locate and read skills that govern record authoring.
Install
Copy the command and run it in a terminal.
npx skills add https:/After installing, type/software-factoryin Claude Code to confirm it loaded.
Bundled files
References and scripts installed along with SKILL.md.
execution/execute-work-order.mdreview.mdwriting-implementation-plans.mdexecution/scripts/checklist-template.mdcontext-template.mdimplementation-plan-template.mdinit-wo-execution.shreview-log-template.mdupdate-context-index.shDetails
From the project READMESoftware Factory
Software Factory is an AI-native SDLC method for connecting product intent, technical intent, and implementation work in one traceable workflow.
This skill equips you to use the Software Factory MCP effectively and guides you through reliable, traceable Work Order execution in this repository.
Records
- The Knowledge Base holds the project's documents and files, such as research, meeting notes, and design artifacts.
- Requirements describe the system from an external perspective: what it must do for its users and why.
- Blueprints describe the system from an internal perspective: its components, contracts, and architecture.
- Work Orders describe delivery: the implementable scope of a change, its exclusions, and the requirements and blueprints it connects to.
- Feedback is requests and reports submitted by the product's users.
- Themes group feedback items that share a common root cause, so they can be triaged together and addressed in one Work Order.
Record authoring
Before you create or change a Knowledge Base document, requirement, blueprint, Work Order, feedback item, or theme, read the project's current writing rules for that kind of record through the Software Factory MCP. Call list_skills to find the writing rules for that kind of record, then call read_skill on that skill and read any child skill it points you to. Write the record exactly as those rules describe. If the Software Factory MCP is unavailable, stop and tell the user.
During implementation, read every referenced Blueprint through the Software Factory MCP before coding, including @… mentions and links.
Routing
| Task | Read |
|---|---|
| Executing one Work Order | execution/execute-work-order.md |
| Executing multiple Work Orders | execution/execute-work-order.md |
| Writing an implementation plan during execution | execution/writing-implementation-plans.md |
| Running the review | execution/review.md |
| Initializing an execution directory | execution/scripts/init-wo-execution.sh |
| Updating execution context | execution/scripts/update-context-index.sh |
| Creating or changing a Software Factory record | Project skills, through the Software Factory MCP |
Work Order Execution
Follow the execution process for every Work Order. Finish every checklist item in one of two states: checked complete with [x], or marked [SKIP] with a skip reason. An unchecked item is an execution failure.
Follow execution/execute-work-order.md for single Work Orders and multi-Work-Order queues, and read the related files in execution/ when that guide routes to them. The execution files record how the work was done; the Software Factory MCP writing rules govern the records themselves.
The checklist is a living harness-engineering artifact. Teams evolve it with the exact commands, checks, screenshots, migrations, fixtures, seed data, CI gates, and review rituals that make agentic programming reliable in their codebase.
Commit, push, open a pull request, or merge only when the user or the repository workflow asks for it.