AgentSkillsCN

plan-agile-planning

创建史诗故事、用户故事，并估算点数，以实现可预测的交付。

SKILL.md
--- frontmatter
name: plan-agile-planning
description: Create epics, user stories, and estimate points for predictable delivery.

Agile Planning Skill

What it does

  • •Translates new requirements into epics and user stories mapped to the active project.
  • •Defines acceptance criteria and estimation (story points/t-shirt sizes) consistent with team velocity.
  • •Produces a prioritized backlog aligned to outcomes and logs decisions in memory.

When to use

  • •Before starting a new initiative or feature set.
  • •When backlog is unclear or lacks acceptance criteria or estimates.
  • •When stakeholders need scope clarity and effort signals.

Key operations

  • •Identify active project (projects/current-project.txt or PROJECT.md status) and review PROJECT.md, backlog/epics.json, backlog/user-stories.json, current-sprint.json.
  • •Identify outcomes and slice into epics and stories with INVEST acceptance criteria.
  • •Estimate effort (points) and dependencies; flag risks/assumptions; check against historical velocity in sprint-history.md.
  • •Prioritize using value vs. effort and constraints; propose sprint placement respecting capacity guidance from AGENTS.md.
  • •Log rationale and decisions in memory/YYYY-MM-DD.md.

Output files

  • •projects/<ACTIVE>/backlog/epics.json — epics with goals, scope, risks, dependencies.
  • •projects/<ACTIVE>/backlog/user-stories.json — prioritized user stories with acceptance criteria and estimates.
  • •projects/<ACTIVE>/backlog/backlog-index.md — human-readable summary.
  • •projects/<ACTIVE>/backlog/assumptions-risks.md — assumptions, constraints, risks, mitigations.
  • •projects/<ACTIVE>/backlog/estimation-log.md — estimation rationale and calibration notes.
  • •memory/YYYY-MM-DD.md — decisions, rationale, sprint placement notes.

Example workflow

  1. •Run session init (AGENTS.md checklist); confirm active project.
  2. •Capture product goal and constraints; scan backlog for overlaps.
  3. •Draft epics and slice into stories with acceptance criteria.
  4. •Estimate each story (points); note dependencies and risks; check velocity.
  5. •Prioritize, propose sprint placement, update backlog files; log decisions in memory/YYYY-MM-DD.md.

Best practices

  • •Make stories INVEST and testable; acceptance criteria are observable.
  • •Calibrate points with recent completed stories; avoid single-person estimates.
  • •Keep epics outcome-oriented; avoid solution-locking too early.
  • •Note risks and unknowns explicitly; create spikes for research.
  • •Respect capacity guidance: planned points ≤ velocity × 1.1 and reserve 10–20% buffer.

Related resources

  • •AGENTS.md (session flow, safety rules)
  • •SKILLS.md, MEMORY.md, memory/YYYY-MM-DD.md
  • •projects/<ACTIVE>/PROJECT.md, backlog/epics.json, backlog/user-stories.json, backlog/backlog-index.md
  • •Definition of Ready/Done; team estimation guide (Planning Poker)