Content
81%Weight 40%Scale 1-5Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
A tightly written, executable destructive-workflow skill with strong validation checkpoints and feedback loops throughout Steps 0–5. The main weakness is minor: referenced helper scripts live outside this skill's bundle.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Mostly lean, instruction-focused prose with concrete commands and no padding about concepts Claude already knows; the occasional rationale notes (e.g., '顺序说明') are justified for a destructive workflow but could be trimmed slightly. | 4 / 5 |
Actionability | Concrete, mostly executable commands throughout (git branch --show-current, git worktree list, git worktree remove [--force], git branch -d) plus full MCP CallMcpTool argument blocks, but placeholders (<TARGET_ROOT>, <issue号>) and conditional handling ('除非…按提示处理', '必要时加') leave minor gaps versus copy-paste readiness. | 4 / 5 |
Workflow Clarity | Steps 0–5 are clearly sequenced with explicit validation checkpoints for a destructive operation: master-branch guard, TARGET_ROOT != MAIN_ROOT check, dirty-tree consent, fetch/merge then git branch --merged, AskQuestion before deleting unmerged branches, and issue-state check before closing — plus feedback loops for -d refusal and MCP failure. | 5 / 5 |
Progressive Disclosure | Well-organized into clear sections (固定参数, 硬性约束, 工作流, 注意事项) as a focused single-file workflow, but it references external scripts (.cursor/skills/github-worktree/scripts/git-worktree.ps1/.sh) that are not present in the local bundle, a minor navigation/organization gap. | 4 / 5 |
Total | 17 / 20 Passed |