AgentSkillsCN

plan-story

通过键值（如NPT-100）获取Jira卡片，并制定详尽的实施计划。在开始一项新故事的工作时使用。

SKILL.md
--- frontmatter
name: plan-story
description: Fetch a Jira card by key (e.g. NPT-100) and produce a detailed implementation plan. Use this when starting work on a new story.
allowed-tools: [mcp__atlassian__getAccessibleAtlassianResources, mcp__atlassian__getJiraIssue, Read, Glob, Grep, Task, EnterPlanMode]

Given a Jira issue key passed as an argument (e.g. NPT-100), fetch the story details and produce a full implementation plan.

Steps

  1. •

    Fetch the Jira issue. Call getAccessibleAtlassianResources to obtain the cloud ID, then call getJiraIssue with the issue key provided as the argument. Extract the summary, description, status, and any linked issues.

  2. •

    Summarize the requirements. Parse the Jira description into a clear list of requirements, acceptance criteria, assumptions, and out-of-scope items. Present this summary to the user so they can confirm understanding before planning.

  3. •

    Explore the codebase. Based on the requirements, use the Explore agent (Task tool with subagent_type=Explore) to investigate:

    • •Existing entities, DTOs, controllers, and EF business-logic classes related to the feature
    • •Similar features already implemented (e.g. if the story is about OVTA, look at how Projects or TreatmentBMPs were built end-to-end)
    • •Database tables, lookup tables, and seed data relevant to the domain
    • •Frontend components, routes, and services for the related area
    • •Any generated files that will need regeneration after backend changes
  4. •

    Enter plan mode. Call EnterPlanMode and write a detailed, step-by-step implementation plan covering:

    Plan structure

    • •Database layer — new tables, columns, lookup tables, seed data, foreign keys, views (including vGeoServer* views if spatial)
    • •EF Models layer — new entities (will be scaffolded), extension methods, business logic classes in Neptune.EFModels/Entities/, validation
    • •DTO layer — new DTOs in Neptune.Models/DataTransferObjects/
    • •API layer — new controllers, routes, authorization attributes ([AdminFeature], [JurisdictionEditFeature], [UserViewFeature], [SitkaAdminFeature]), any new permissions/rights
    • •Tests — integration/unit tests in Neptune.Tests (placed here so the user can run them while frontend work proceeds)
    • •Code generation — run /codegen to regenerate TypeScript models/services
    • •Frontend layer — new standalone components, lazy-loaded routes, services, ag-Grid grids, reactive forms, modals

    For each step, reference specific existing files as templates to follow (e.g. "Model after TreatmentBMPController.cs") and call out the exact file paths where new code should go.

  5. •

    Present the plan for user approval via ExitPlanMode.

Notes

  • •The cloud ID for the Jira instance is obtained dynamically via getAccessibleAtlassianResources.
  • •Always follow existing patterns in the codebase — find the closest analogous feature and replicate its structure.
  • •The plan should be detailed enough that each step can be executed independently.
  • •Reference the CLAUDE.md architecture patterns (authorization attributes, jurisdiction-scoped routes, Observable + async pipe, standalone components, etc.).
  • •Do NOT start writing code — this skill only produces a plan.