Generated from cat-harness/skills/process/workflow/release-epic-planning.md — do not edit here.

✎ Edit this page’s source

Release epic planning

A release epic groups all the work for a specific release version. This skill governs how to create, structure, and track them.

Epic structure

The beans hierarchy for a release:

epic: "Release v0.2.0"
├── feature: "CRDM Phase 4b prioritization"
│   ├── task: "Add Phase 4b to crdm-requirements-workflow.md"
│   └── task: "Update phase table"
├── feature: "Release lifecycle skills"
│   ├── task: "Create release-lifecycle.md"
│   ├── task: "Create code-review-process.md"
│   └── task: "Create branch-freshness.md"
└── bug: "Fix 7 BPMN isMarkerVisible attributes"
    └── task: "Add isMarkerVisible to 5 files"

Bean types in a release context

TypeRole in releaseGranularity
epicThe release itself1 per release
featureA capability area / user story1 per feature
taskA single PR-sized implementation unit1 per PR
bugA defect fix included in the release1 per fix
milestoneA session-scoped work containerNot release-scoped

Creating the hierarchy

# 1. Create the release epic
beans create "Release v0.2.0" --type epic

# 2. Create features under it
beans create "CRDM Phase 4b prioritization" --type feature
beans update <feature-id> --parent <epic-id>

# 3. Create tasks under features
beans create "Add Phase 4b to workflow" --type task
beans update <task-id> --parent <feature-id>

Always check before creating — beans create is not idempotent. See todo-manager.md §”Check before you create”.

Ordering within a release

Apply the Phase 4b formalize-first rule from crdm-requirements-workflow.md:

  1. Structural/organizational work first
  2. Process/policy formalization second
  3. Implementation third
  4. Visualization/UI last

Within the epic, number features by dependency order. The first feature completed should not depend on any later feature.

Release readiness

Tracking readiness

The beans visualizer (<base>/beans/) shows open beans grouped by epic. A release is ready when its epic shows 0 open beans.

Readiness checklist

Before moving from draft to pre-release:

  • All must-have features resolved (all child beans completed)
  • All should-have features either resolved or deferred to next release
  • CI green on main (bun run cat gates)
  • All changed BPMN diagrams re-rendered
  • All changed docs pages regenerated
  • Changelog written (see release-lifecycle.md)
  • BA has reviewed accumulated changes

Deferring to next release

If a should-have or nice-to-have feature won’t make the release:

# Move the feature bean to the next release epic
beans update <feature-id> --parent <next-epic-id>

Record the deferral reason in the bean body:

beans update <feature-id> --body-append "Deferred to v0.3.0: not blocking v0.2.0"

GitHub milestone integration (optional)

A GitHub milestone can mirror the release epic for stakeholder visibility:

# Create milestone
gh api repos/{owner}/{repo}/milestones --method POST \
  -f title="v0.2.0" -f description="Release v0.2.0"

# Assign PRs to milestone
gh pr edit <number> --milestone "v0.2.0"

The milestone’s completion percentage gives stakeholders a progress view without needing access to the beans store.

Cross-domain applicability

DomainEpic titleFeatures
Code“Release v0.2.0”Skills, schemas, pipeline changes
Paper“Publication: Chapter 3 revision”Section rewrites, figure updates
Project“Deliverable: Phase 1 report”Analysis sections, appendices

The hierarchy and readiness tracking are identical.

Rules

  • One epic per release — if scope creeps, split the epic
  • Features are independently reviewable — each can be a PR
  • Tasks are PR-sized — one task, one PR, one review cycle
  • Never close an epic without BA approval — the epic is the release
  • Deferred features move to the next epic — they are not deleted

Cross-references

Processes that run this skill

processstep(s) that name it
Phase 1: PlanningDraft project roadmap, backlog, and sprint cadence

folio-assistant — code under the Apache License 2.0, documentation under CC BY 3.0. Generated schema and API references are rebuilt from source on every push to main.

deployed dev