| name | plan-delivery-track |
| description | Use to decompose product or design context into a delivery tracker and lightweight story briefs. Accepts a PRD plus technical solution, a technical solution alone, or explicit backlog/design context when enough scope, outcomes, sequencing, and validation expectations exist. For complex technical work, require a technical solution or equivalent external design context before writing tracker stories. Produces docs/tracks/<track>/README.md plus docs/tracks/<track>/stories/<ID>.md briefs that cite PRD criteria or context-derived outcomes and technical solution sections or external context labels. Story briefs are not implementation-ready; implement-next creates the detailed technical story spec and implementation plan before code. |
| argument-hint | [prd-slug or notes] |
| arguments | prd_slug_or_notes |
| user-invocable | true |
Plan a delivery tracker
Turn agreed product/design input into delivery sequencing: a tracker of bounded stories and
lightweight story briefs. The input can be a PRD plus technical solution, a technical solution alone, or explicit backlog/design context when enough scope exists. This skill does not write
detailed technical story specs and does not write implementation plans.
Where this sits
define-product (PRD) -> design-technical-solution (high-level technical solution when needed)
-> plan-delivery-track (tracker + story briefs) -> implement-next (detailed technical story
spec -> implementation plan -> code) -> workflow-autopilot or human-driven story execution.
The workflow altitude is:
PRD -> Technical solution -> Delivery tracker + story briefs -> Detailed story spec -> Implementation plan -> Code
Bundled references (load before producing each artifact)
| Reference | Use for |
|---|
references/prd-contract.md | PRD input and acceptance-criteria IDs |
references/technical-solution-contract.md | Technical solution gate and required input shape |
references/tracker-contract.md | Tracker README contract and 9-column status matrix |
references/story-brief-contract.md | Story brief contract |
references/templates/tracker/tracker-readme-template.md | Tracker README template |
references/templates/story-brief-template.md | Story brief template |
Config
Read .workflow/config.yaml if present; otherwise use defaults. Keys used:
paths.prdsDir (default docs/prds), paths.tracksDir (docs/tracks), statuses.*, and
tracker.idPattern (default ^[A-Z]{2,}[0-9]+$). This skill writes no config.
The recipe
Step 1 - Source context gate
Resolve paths.prdsDir. Locate the PRD for this work (<prdsDir>/<slug>/, conforming to
references/prd-contract.md) when a PRD slug or path is supplied. Read 08-acceptance-criteria.md;
every story brief maps to one or more PRD criteria when PRD criteria exist.
If no conforming PRD exists, continue only when the user supplied a technical solution alone or
explicit backlog/design context with enough product scope, acceptance outcomes, sequencing
constraints, and validation expectations to create responsible tracker rows. Record that assumption
in the tracker and each story brief. If that context is missing or contradictory, stop and tell the
user to run /define-product first.
Step 2 - Technical solution gate
Classify whether this is a complex technical PRD. Technical solution is required when the PRD
implies any new backend modules, shared services, database schema/query changes,
AI prompts/triggers/tools, observability/events/metrics, migration/deploy surfaces,
security/privacy boundaries, multi-system integration, or multiple implementation stories whose
sequencing depends on technical design.
If technical solution is required and missing, stop and tell the user to run
/design-technical-solution first unless explicit external design context already covers the same
high-level how. Do not decompose complex technical work directly from a thin PRD.
If the technical solution is present, read <prdsDir>/<slug>/technical-solution.md and confirm it
conforms to references/technical-solution-contract.md. Use it as the source for story boundaries,
sequencing, validation expectations, file-contention constraints, and required section citations.
For technical-solution-only or backlog/design-context planning, cite the supplied document headings
or context labels in the story brief's technical solution sections.
Step 3 - Audit reality
Read repo contract docs (AGENTS.md, CLAUDE.md, GEMINI.md, architecture docs), source roots,
and existing trackers under paths.tracksDir. Repo reality wins if it contradicts the PRD or
technical solution; flag the mismatch rather than copying stale assumptions into story briefs.
Step 4 - Reserve the ID prefix
Read <tracksDir>/README.md for the prefix registry. Pick a 2+-letter prefix that is disjoint from
every reserved prefix and matches tracker.idPattern.
Step 5 - Identify stories
A story is one future implement-next session that can create a detailed technical story spec,
write an implementation plan, code, verify, and open one PR. Aim for 5-25 stories.
For each story, define PRD criteria and technical solution sections, dependencies, scope boundary,
candidate surfaces, validation expectations, and open technical questions that the detailed spec
must resolve.
Add an Assumptions and blockers pass before writing: record safe assumptions and ask only
questions that block a valid tracker or story brief. Add an Artifact boundaries pass: confirm the
tracker owns delivery slicing and status, story briefs are lightweight and not implementation-ready,
detailed technical story specs own exact implementation design, implementation plans own execution
steps, and runtime artifacts own execution evidence.
Step 6 - Write tracker README
Write <tracksDir>/<track>/README.md from tracker-readme-template.md. Keep the 9-column status
matrix exactly. For new trackers, the Spec column links to the story brief, for example
[brief](./stories/<ID>.md). Leave Plan as —; implement-next creates the implementation
plan later.
Step 7 - Write story briefs
Write one lightweight story brief per tracker row at:
<tracksDir>/<track>/stories/<ID>.md
Use references/templates/story-brief-template.md. Each brief must include PRD criteria or
context-derived outcomes, technical solution sections or external context citations, dependencies,
scope boundary, assumptions and blockers, artifact boundaries, candidate surfaces, validation
expectations, open technical questions, and the exact note:
not implementation-ready; create a detailed technical story spec before plan/code
Remember: story briefs are not implementation-ready. Do not write detailed technical story specs or
implementation plans from this skill.
Step 8 - Register the track
Add the new track row and reserved prefix to <tracksDir>/README.md. A tracker that is not
registered is invisible to orchestrated selection and risks prefix collision.
Step 9 - Summarize and hand off
List the tracker README and story briefs written. Cross-check that every matrix row links a story
brief and every brief links back to the tracker, PRD, and technical solution when present. Point the
user at /implement-next for detailed story-spec creation, implementation planning, and code.
Do not auto-commit and do not invoke implementation skills.
Backward compatibility
Existing trackers that link a detailed spec directly (rather than a story brief) remain valid. New
trackers should link story briefs under <tracksDir>/<track>/stories/.
Anti-patterns
- Writing detailed technical story specs in
plan-delivery-track.
- Writing implementation plans before
implement-next.
- Treating a story brief as implementation-ready.
- Adding tracker columns or statuses.
- Skipping the Technical solution gate for complex technical work.