Skip to content
Skill Authoring
Skill

/factory-product

Use when orchestrating a product lifecycle (specify/init → clarify → implement → analyze) to generate or verify PRD.md.

BOOST
From plugin
adlc-team-skills
14345 skills
Install
$ npx -y skills add tikalk/adlc-team-skills --skill factory-product --agent claude-code

How it fires

How this skill gets triggered: by you, by Claude, or both.

  • Fires itselfAuto-invocation. Claude auto-loads it when your prompt matches the work.Auto-invocation is when the right skill fires by itself at the right moment, driven by a FLOW.md router and a hook, instead of you invoking it by name. It is the difference between a skill being installed and a skill actually getting used.Read the full definition →
  • You can call itInvoke it directly when you want it.
  • Slash command/factory-product

Context preview

The summary Claude sees to decide when to auto-load this skill.

Use when orchestrating a product lifecycle (specify/init → clarify → implement → analyze) to generate or verify PRD.md.

SKILL.md

factory-product.SKILL.md
name: factory-product
description: Use when orchestrating a product lifecycle (specify/init → clarify → implement → analyze) to generate or verify PRD.md.
disable-model-invocation: true

factory-product

What this skill does

`factory-product` orchestrates the product-decisions lifecycle. It coordinates individual product-related skills (`product-init`, `product-specify`, `product-clarify`, `product-implement`, `product-analyze`) to maintain a consistent `PRD.md` at `docs/adlc/product/PRD.md` (the shared layout).

It operates as a **Kind-A DAG orchestrator** in alignment with the shared executor engine contract in `factory-mission/references/executor.md`.

---

When to use

  • You want to bootstrap, refine, or analyze a product's requirements and decisions (PDRs) end-to-end.
  • You want a guided, accept-gated, and verified product definition process with a resume checkpoint.

**When NOT to use**:

  • For software implementation task execution (use `factory-mission` instead).
  • For architecture decision records (use `factory-architect` instead).

---

Lifecycle DAG & Step Resolution

`factory-product` implements a **fixed named-skill DAG** (`fixed` step resolution):

Greenfield Route (default on empty project)

1. **`specify`** (`generate` phase) -> Invoke `product-specify` to collaboratively capture PDR drafts in `.adlc/drafts/pdr/`. 2. **`clarify`⭐** (`clarify` phase) -> Invoke `product-clarify` to run interactive quality checks and mark PDRs Accepted (human sign-off gate). 3. **`implement`** (`build` phase) -> Invoke `product-implement` to compile accepted PDRs into `docs/adlc/product/PRD.md` and promote them to `docs/adlc/memory/pdr/`. 4. **`analyze`** (`analyze` phase) -> Invoke `product-analyze` to verify PRD/PDR consistency and output a severity-ranked report.

Brownfield Route (default if code exists but no PDRs)

1. **`init`** (`generate` phase) -> Invoke `product-init` to reverse-engineer draft PDRs from the existing codebase. 2. **`clarify`⭐** → **`implement`** → **`analyze`** (same as Greenfield).

Refresh Route (default if PRD.md and memory PDRs already exist)

1. **`analyze`** (`analyze` phase) -> Run product-analyze first to detect drift. 2. **`clarify`⭐** → **`implement`** → **`analyze`** (drift-correction cycle).

---

Shared Executor Overrides

`factory-product` overrides the shared executor engine primitives as follows:

1. **Publish Target**: Fixed to `local`. Outputs are written to `docs/adlc/memory/pdr/` and `docs/adlc/product/PRD.md`. If tracker-integrated, a `tracker` completion summary comment is also posted. 2. **Output Types**: Steps use the following `output_type` assignments:

  • `specify`/`init` → `draft` (PDR drafts stay in `.adlc/drafts/pdr/`, not published to comment bus)
  • `clarify`⭐ → `decision` (accepted/rejected PDR list published to comment bus)
  • `implement` → `artifact-ref` (PRD.md path reference published, content stays on disk)
  • `analyze` → `findings` (severity-ranked consistency report published to comment bus)

3. **Correction Loop**: If `analyze` returns `CRITICAL` or `HIGH` consistency errors, the executor routes back to `clarify` with the analyze marker in `reads_from`. `clarify` reads the findings from the PR/MR/issue comment bus (or local fallback). This loop is bounded by `max_corrections` (default 2); if exceeded, the orchestrator halts for human review. 4. **Supervision Default**: `hybrid`. A human gate is hard-enforced at `clarify`⭐ (for PDR approvals) and at final `PRD.md` review. 5. **Pre-flight Check**: Verifies that the `product-*` lifecycle skills are installed in the workspace before beginning Phase 0.

Read more
Ships withadlc-team-skills

Agent skills that give coding agents your team's context at session start, so they stop working like strangers.

Get the whole plugin

Other skills on adlc-team-skills.