backlog-prioritization
Use when ordering board tasks - set priority/blocked_by/deploy_depends_on by severity and dependency, split MoSCoW within a feature, and use RICE/ICE only for a genuine "what next" call across several features
- License 许可
- Apache-2.0
- 最近更新
- 2026-11-15
- Stars
- ★ 122
- Forks
- 18
01 前言 Front Matter
- name
- backlog-prioritization
- source
- phuryn/pm-skills prioritization-frameworks (MIT), adapted
- category
- pm
- description
- Use when ordering board tasks - set priority/blocked_by/deploy_depends_on by severity and dependency, split MoSCoW within a feature, and use RICE/ICE only for a genuine "what next" call across several features
03 SKILL.md 正文
Backlog Prioritization
Overview
Prioritization decides what agents pick up next. TaskTrooper is a local, single-user product with no usage data, so a scored framework with invented numbers (fake "Reach 200") is theatre. Order by the levers the board actually enforces, and fall back to an explicit estimate only when several same-level features genuinely need ranking.
Core principle: priority, blocked_by, deploy_depends_on and backlog/todo are the real order. Score only when they tie.
The real levers
| Lever | What it controls |
|---|---|
priority (critical / high / medium / low) |
The only persisted order field; sorts list_ready_tasks critical-first, then by age. |
blocked_by |
The enforced START order — a task parks in blocked while any listed task is open, and is picked up automatically when the last one lands. |
deploy_depends_on |
The enforced SHIP order — this task is not released until the listed ones are live. |
| backlog vs todo | The human gate — nothing in backlog is picked up; moving to todo starts it. |
Priority ladder
- critical — data loss or corruption, a security exposure, the product unusable, a broken release.
- high — a core flow broken, a regression of shipped behaviour, an analiz that unblocks two or more tasks, a stated deadline.
- medium — default for a requested feature.
- low — cosmetic, nice-to-have, a Should/Could split-off from MoSCoW.
Within a level: the stakeholder's stated order, then unblockers (an analiz or a backend task other work depends on), then smaller work first.
MoSCoW inside a feature
When scoping one feature or task:
- Must → this task's
acceptance_criteria. - Should / Could → separate backlog tasks, not folded into this one's criteria — one task per repository and role holding all of that feature's Should/Could items, not one task per item.
- Won't → the
description's Out of Scope line, with the reason.
RICE/ICE — only for an explicit "what's next" call
Use a lightweight score only when the stakeholder asks "what should we do next" across three or more same-priority-level features, and label every factor as an estimate with its basis — never invent a reach number with no usage data behind it.
score = (Impact × Confidence) ÷ Effort (ICE; drop Reach — there's no multi-user traffic to count). Impact and Confidence are 1–3 with a one-line reason each; Effort is developer-days from the architect's estimate if one exists, otherwise your own rough guess labelled as such.
Worked Example
Three candidates in backlog:
- A data-loss bug on task delete →
priority: critical,todoimmediately — ladder override, no scoring needed. - An analiz for reporting that unblocks three planned tasks →
priority: high(unblocker),todo. - A new dark-mode toggle the stakeholder mentioned once →
priority: medium, stays inbackloguntil they confirm it's next.
If the stakeholder then asks "dark mode, bulk delete, or CSV export — what first?": ICE each in one line (dark mode: Impact 1/low because cosmetic, Confidence 3, Effort 2 → ICE 1.5; bulk delete: Impact 2, Confidence 3, Effort 1 → ICE 6; export: Impact 2, Confidence 2, Effort 3 → ICE 1.3) and recommend bulk delete first, with the basis stated, not just the number.
Common Mistakes
- Scoring every backlog item with RICE instead of using
priority/blocked_bydirectly. - Inventing a reach/impact number with no basis and presenting it as data.
- A Should/Could item folded into a Must task's criteria instead of split off.
- Holding work in
backlogto fake an order thatblocked_byshould enforce.
Red Flags
- A known data-loss bug sitting below a feature in
priority. - An unblocking analiz left at
mediumor inbacklog. - A RICE score with a "Reach" number nobody can source.
04 同仓库更多 Skills
visual-craft
暂无简介
ux-copy
Use when you write any words a user reads in a mockup or a hand-off - content language and voice, buttons, labels, errors, empty states, confirmations, numbers and dates, and the longest-string check
screen-mockup-html
Use when you draw screens as the design task's HTML mockup - ONE canvas document holding every screen and every variant side by side, the sanitizer-safe recipe, phone and wide frames, the states grid, static control states, and the loopback self-check with screenshots at 375, 768 and 1440 attached to the task
performance-awareness
How the performance score moves for your role. Use when the score context shows a declining trend, or before handing off work you have not self-checked against every criterion.
design-variants
Use when a screen design needs variants - how many, variants as side-by-side rows on the one canvas, what makes two variants genuinely different, recommending one, how the human's choice is read, and narrowing to the chosen one in revision
design-system-authoring
Use when you derive, change or extend a design system - where tokens live per stack, the DTCG token tree, role-based names for both themes, base vs layer vs drift, the DESIGN.md section order, the inventory format, the contrast check and the lint every proposal returns