ai-dlc
Skill1 starsChanged 3 months agoMIT-0View on GitHub
Run the full AI-DLC inception pipeline for a task — story → refine → design → tasks — with a human gate between each phase
In this repository8 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- commitSkill
- design-docSkill
- pr-draftSkill
- spec-designSkill
- story-newSkill
- story-refineSkill
- tasksSkill
The file
We split this skill into the rules it gives an agent
1 rule. About 0 are common, 0 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 28
no folder / no story.md → start at phase 1.
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
/ai-dlc — inception pipeline orchestrator
Walks a task through the AI-DLC inception phases end to end, pausing for human validation between each. A thin conductor over the per-phase skills — it does not re-implement them.
Phases (each is a separate skill; run them in order)
- Story →
story-new— draftstory.mdfrom the idea. - Refine →
story-refine— resolve open questions, verify facts, tighten scope. - Design →
design-doc— architecture + task map (+ ADRs if warranted). - Tasks →
tasks— expand the task map intotasks/(API surface per unit, not full code).
pr-draft is deliberately not part of this orchestrator: it runs after the building phase, once real code exists on the branch. Inception stops at the task interfaces.
Spec-driven entry. When the work starts from a factory spec (specs/<name>.yaml) rather than a rough idea, phases 1–2 are already satisfied — a spec is refined requirements. Enter at phase 3 via spec-design (spec → design-doc.md) instead of design-doc, then continue at phase 4 (tasks). The spec, not a story.md, is the source of truth; re-run spec-design when the spec changes.
Input
An idea and optional slug, e.g. /ai-dlc weather-tool "MCP tool that fetches current weather". If resuming an existing story, pass just the slug and this skill picks up at the first incomplete phase.
How to run
- Determine the starting phase. First, if a
specs/<slug>.yamlexists (spec-driven work), enter at phase 3 viaspec-designand skip phases 1–2 entirely. Otherwise check.local/user-stories/<slug>/for existing artifacts:- no folder / no
story.md→ start at phase 1. story.mdpresent but has⏳ open:items → start at phase 2.- refined
story.md, nodesign-doc.md→ phase 3. design-doc.mdpresent, notasks/→ phase 4. State which phase you're starting at and why.
- no folder / no
- Run each phase by invoking its skill (the Skill tool), in order, passing the slug.
- Gate between phases. After a phase produces its artifact and the user approves it, ask the user to confirm before advancing to the next phase. If the user wants to stop after any phase, stop — each artifact stands on its own and the pipeline is resumable later.
- Never skip a gate or auto-advance past an unapproved artifact. The value of AI-DLC here is the human validation between steps; preserve it. A phase's output must be approved before the next phase consumes it.
- After phase 4, tell the user inception is complete and the tasks are ready for the building phase;
/pr-draft <slug>comes later, after code is written.
Principles
- Thin orchestration only — defer all phase logic to the per-phase skills so behavior stays consistent whether a phase is run standalone or via
/ai-dlc. - Resumable and re-runnable: any phase can be re-invoked to revise its artifact without redoing the others.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

