Content
93%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 lean, highly actionable instruction skill with concrete tool names, parameters, and a clear sequenced workflow including a verification step. The only gap is the absence of an explicit retry-on-failure feedback loop in the verification step.
Suggestions
Add an explicit feedback loop after the dashboard-insights-run verification step (e.g., 'if a tile returns no data, adjust its query or remove it before finalizing').
Clarify the dashboard-insights-run verification as a checkpoint with a pass/fail branch rather than a one-line confirm-and-summarize.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Lean and efficient; assumes Claude knows PostHog/MCP, names concrete tools and params without padding, with only one mildly editorial line ('a good bespoke dashboard beats a mismatched template every time'). | 5 / 5 |
Actionability | Fully actionable: names every relevant MCP tool with key params (search, scope, tiles, layouts.sm, x/y/w/h) and gives concrete layout rules (12-column grid, widths 1-12, 8+4/6+6) covering the common create/update cases. | 5 / 5 |
Workflow Clarity | Clear sequenced sections (create vs update → templates → select insights → assemble → verify) with a verification checkpoint via dashboard-insights-run and an explicit guardrail for the destructive update behavior, but lacks an explicit validate→fix→retry feedback loop. | 4 / 5 |
Progressive Disclosure | Well-organized single-purpose skill with clear section headers, a 'When not to use this' boundary, and one-level 'Related skills' pointers; no bundle files are needed and none are missing. | 5 / 5 |
Total | 19 / 20 Passed |