Content
53%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.
The body is a well-structured orchestrator index with clear phase sequencing and copy-paste invocation prompts, but it stops at the directory level: the Actions lists are generic placeholders with no executable specifics, and the only validation artifact (the quality-gates checklist) is not wired into the workflow. The malformed testing-pyramid ASCII diagram (its code fence closes mid-diagram) is a minor quality blemish. It could score much higher by moving per-phase detail into reference files and adding explicit gate-enforcement steps.
Suggestions
Replace the generic numbered 'Actions' lists with executable specifics — actual commands, config snippets, or example test code per phase (e.g., a sample pytest/Jest/Playwright invocation and a coverage-threshold CI gate) instead of 'Define testing strategy' / 'Measure coverage'.
Wire the Quality Gates Checklist into the workflow as explicit checkpoints with failure handling: after each phase, state which gate to run, the exact command, and the fix-and-retry loop when a gate fails.
Move per-phase detail (framework setup, CI integration, visual regression config) into one-level-deep reference files (e.g., references/unit-testing.md) to cut the repetitive inline scaffolding, and fix the broken testing-pyramid code fence.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body avoids explaining concepts Claude already knows (no 'what is unit testing' padding), but the eight-phase template is highly repetitive: each phase restates the same 'Skills to Invoke / Actions / Copy-Paste Prompts' scaffolding, and many prompts trivially restate the skill name ('Use @find-bugs to detect bugs in code'). More than minor trimming is needed, so it sits at anchor 3 rather than anchor 4. | 3 / 5 |
Actionability | The copy-paste prompts are concrete, literal invocations ('Use @playwright-skill to create E2E tests'), which is real executable guidance for an orchestrator skill. However, all 40+ 'Actions' items are vague directives with no commands, code, or specifics ('Define testing strategy', 'Configure API mocks', 'Measure coverage'), and no example demonstrates an actual test or command — concrete guidance present but incomplete, matching anchor 3. | 3 / 5 |
Workflow Clarity | The eight phases are clearly sequenced (Strategy → Unit → Integration → E2E → Browser → Performance → Review → Gates), but validation checkpoints are missing from the flow itself: the 'Quality Gates Checklist' is a standalone unchecked list not tied to any phase, there are no validate → fix → retry loops, and nothing tells Claude what to do when a gate fails. Anchor 3 ('checkpoints missing or implicit') fits better than anchor 4. | 3 / 5 |
Progressive Disclosure | Structure is good: consistent per-phase headers, a clear overview, a when-to-use section, and clearly signaled @-skill references make navigation easy. At ~230 lines with all eight phases inlined, per-phase detail (e.g., framework setup, CI config) belongs in one-level-deep reference files, which keeps it below anchor 5's 'content appropriately split' — no bundle files exist, so everything is in SKILL.md. | 4 / 5 |
Total | 13 / 20 Passed |