abd-thin-slicing

name: abd-thin-slicing catalog_garden_tier: practice catalog_garden_order: 20 catalogue_one_liner: >- Decide what to deliver first and why — so the team ships value in the smallest useful increments. description: >- Produce thin-sliced delivery increments with priority order, outcomes, and story groupings. Use when planning releases or deciding what to deliver first after story mapping. context-perspective: stories context-role: support context-fidelity:

• level: discovery mode: slice-ordering

abd-thin-slicing

Purpose

Decide what to deliver first and why — so the team ships value in the smallest useful increments.

Grill prompts

Read common/grill-me-with-practice-skill.md before grilling.

Before generating, surface these:

• Spine vs optional — Which stories must be delivered together to show an end-to-end path, and which can follow later without blocking value?
• Vertical not horizontal — Are you slicing by user-visible capability, or by technical layer — and would a stakeholder recognise your increment names?
• Value assumption — What makes you believe this increment is the smallest useful thing, rather than a comfortable batch?
• Dependency trap — Which cross-epic dependencies are you hiding inside an increment instead of making them visible?
• Ordering rationale — Why does increment N come before increment N+1 — is it risk, learning, or just the order you thought of them?

Output file

Deliverables folder: see ../common/skill-rule-workflow.md — Output file resolution.

File name: thin-slicing.md. Add a <name>- prefix only when disambiguation is needed.

Diagram workflow

Produces docs/stories/thin-slicing.drawio from story-graph.json. Requires story-graph.json to exist (built by story-graph-ops). Must exist before the cell is marked done.

python drawio_story_sync_cli.py render \
  --mode thin-slicing \
  --graph docs/stories/story-graph.json \
  --out   docs/stories/thin-slicing.drawio

Agent Instructions

Follow ../common/skill-rule-workflow.md — read-gates, output file resolution, and the per-rule verdict format are defined there.

1. Read context

Read these files:

• reference/concepts.md — definitions of increments, thin slicing, and vertical vs horizontal slices.
• reference/examples.md — output shape examples and weak patterns to avoid.

2. Generate

Produce the template:

New files under templates/ later → one deliverable per file.

Workflow steps:

1. Read inputs — story map / graph, PO or tech notes, known risks and dependencies.

2. Mark spine vs optional — mandatory core sequence vs alternates, enhancements, deep error paths (see rules/map-sequential-spine-vs-optional-paths.md).

3. Cut vertical slices — each increment is an end-to-end demonstrable path (even if manual/stubbed); avoid horizontal "finish epic A, then epic B."

4. Name for value — increment titles = stakeholder-visible capability, not phase or stack labels.

5. Pull stories — under each increment, list verb-noun stories in flow order; don't paste the whole map unless asked.

   Story names MUST be copied verbatim from story-map.md / story-graph.json — character-for-character, including every parenthetical qualifier.

   • - Load FX Resource Catalog (FxRepo.data) (correct — exact)
   • - Load FX Resource Catalog (WRONG — trimmed; creates an orphan; scanner will catch this)
   • - System --> Load Crowd from Repository (WRONG — actor prefix; parser stores "System" as the name)

6. Write the template file — fill templates/thin-slicing.md with the increments and stories (italics on domain terms where helpful).

7. Omit maintainer noise — do not copy the template's ## Instructions block into project deliverables.

Before: abd-story-mapping — produces the story map this skill slices. After: abd-story-acceptance-criteria, abd-story-specification — add story detail once priorities are set.

3. Validate

Run the story-name exact-match scanner first:

python skills/abd-thin-slicing/scanners/story-name-exact-match-scanner.py --workspace <path-to-project>

Exit code 1 means story name mismatches — do not proceed until it passes.

Run scanners and emit per-rule verdicts — see ../common/skill-rule-workflow.md § Validate output.

Validate

• Each increment shows a vertical path (input → processing → outcome), not a layer.
• Increment names are stakeholder-visible capabilities, not internal milestones.
• Quality trade-offs are named ("Manual…" → "Automated…"), not hidden.
• Story names in thin-slicing.md are character-for-character identical to the source story map.