AgentSkillsCN

ticket-feature-checklist

当您新增或扩展工单运行时（src/agent_loom/ticket/*）时，应确保 CLI 行为与核心/存储层保持一致，并通过 uv 工具配合 LSP+ruff+针对性测试，进一步提升整体协调性。

SKILL.md
--- frontmatter
name: ticket-feature-checklist
description: Use when adding/expanding ticket runtime (src/agent_loom/ticket/*): keep CLI behavior coherent with core/store, with LSP+ruff+targeted tests via uv.
license: MIT
compatibility: opencode,claude
metadata:
  created_at: "2026-01-31T02:21:31.200Z"
  updated_at: "2026-01-31T02:21:31.200Z"
  version: "1"
<!-- BEGIN:compound:skill-managed -->

When to use

  • •You changed src/agent_loom/ticket/cli.py, src/agent_loom/ticket/core.py, or src/agent_loom/ticket/store.py.
  • •You changed ticket file format / serialization (src/agent_loom/ticket/frontmatter.py).
  • •You added/changed ticket state fields, storage shape, or CLI behavior.

Goal

Keep ticket behavior consistent across CLI, core logic, and persistence; keep invariants testable.

Checklist

  1. •

    Surface area audit

    • •Identify the behavior you changed: CLI parsing/output, core semantics, persistence shape, adapters/API.
    • •Confirm the CLI maps cleanly to core operations (no duplicate business logic in CLI).
  2. •

    Persistence + format invariants

    • •If store schema/format changed, define upgrade/migration expectations.
    • •Ensure read/write paths are symmetric.
    • •If frontmatter/serialization changed, add a round-trip test (load -> write -> reload) to lock the format.
    • •Ensure errors are actionable (what file, what field, what expected shape).
  3. •

    Deterministic output

    • •Keep CLI output stable and parseable where possible.
    • •Use explicit ordering for fields/lists; do not rely on dict/set iteration.
  4. •

    UX contract tests

    • •If output formatting/UX changed, add/update a focused contract test in tests/test_ticket_ux.py.
  5. •

    LSP first

    • •Run lsp_diagnostics on touched files.
    • •Fix errors/warnings before lint/tests.
  6. •

    Lint

    • •Run uv run ruff check ..
  7. •

    Targeted tests

    • •Add/adjust focused tests for the changed behavior (CLI + core + store + format).
    • •Start with uv run pytest tests/test_ticket_ux.py when UX/output is involved.
    • •Otherwise run the smallest relevant subset first, then expand if needed.

Common failure modes

  • •CLI behavior changes but core/store invariants are not covered by tests.
  • •Store changes land without clear handling of existing on-disk data.
  • •Frontmatter/file-format changes land without a round-trip regression test.
  • •Output becomes order-dependent (dict/set iteration) and breaks snapshots or tooling.
<!-- END:compound:skill-managed -->

Manual notes

This section is preserved when the skill is updated. Put human notes, caveats, and exceptions here.