Content
85%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This is a well-structured skill that provides clear, actionable guidance for breaking down changes into task checklists. Its strengths are the concrete examples (good vs bad tasks table), clear multi-step workflow with mode-conditional logic, and appropriate use of external references. Minor verbosity in the phase organization guidelines and some template repetition prevent a perfect conciseness score, but overall it's a strong skill.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill is mostly efficient but includes some redundancy — the phase organization guidelines and task writing rules table, while useful, overlap somewhat. The format templates are helpful but could be slightly tighter. It doesn't explain concepts Claude already knows, which is good. | 2 / 3 |
Actionability | Highly actionable with concrete file paths, specific format templates, a clear table distinguishing good vs bad task examples, and explicit output format. The task writing rules table with ✅/❌ examples is particularly effective for guiding behavior. | 3 / 3 |
Workflow Clarity | Clear 5-step sequential workflow with explicit steps (Load Skills → Analyze Design → Write tasks.md → Persist Artifact → Return Summary). Each step has clear instructions, and the mandatory persistence step is called out explicitly. The mode-conditional branching (engram/openspec/hybrid/none) is well-structured with clear instructions per mode. | 3 / 3 |
Progressive Disclosure | Appropriately references shared skill files (`sdd-phase-common.md` Sections A-D, `openspec-convention.md`) without inlining their content. References are one level deep and clearly signaled. The skill itself is well-organized with headers for each step and collapsible detail where needed. | 3 / 3 |
Total | 11 / 12 Passed |