Content
100%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.
An exemplar skill body: terse, fully executable, with a validated multi-step workflow and clean one-level-deep references. No bundle files are present in this skill's directory, so references are scored as written; they read as cross-skill pointers rather than local bundle files.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Lean and efficient throughout — terse steps, tables, and 'Do not' bullets with no explanation of concepts Claude already knows; even the enterprise-submodule prose is conditional operational logic where every token earns its place. | 5 / 5 |
Actionability | Provides copy-paste-ready commands for the core flow (`git worktree add -b <branch> <worktree-path>`, `git rev-parse --show-toplevel`, `git submodule update --init --depth 1 .source`, `pnpm symlink:submodules`, `move_agent_to_root`) plus a concrete example; the single deferral (rsync env-copy recipe) is an explicit, named progressive-disclosure pointer rather than a guidance gap. | 5 / 5 |
Workflow Clarity | Clear six-step sequence with explicit validation checkpoints — preflight aborts if the target path exists and warns on a dirty tree with user confirmation, enterprise setup degrades gracefully, and a final 'Confirm' checklist covers branch, path, base ref, copied env, and submodule status. | 5 / 5 |
Progressive Disclosure | A concise overview body that delegates detail to well-signaled one-level-deep references ('read `nv-worktree-commands`', 'read `nv-worktree-cleanup`', 'read `enterprise-submodule`'), with content appropriately split across clearly labeled sections (Inputs, Steps, Example, Do not). | 5 / 5 |
Total | 20 / 20 Passed |