Content
88%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 excellent instruction-only skill body: lean, highly actionable, and sequenced with explicit validation, retry, and spend-confirmation checkpoints. The only real gaps are the unspecified argument syntax for the main research tool and the absence of any reference files to offload the report-format and tool-detail content.
Suggestions
Show one example invocation of `research_ai_visibility_prompts` with its actual argument shape (e.g., the keyword parameter), as is already done for `update_project_context`.
Move the report section template and tool semantics ("What to return", "OpenSEO MCP tools") into a one-level-deep reference file to shorten the always-loaded body.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense and operational throughout — tool names, thresholds ("5–15 candidate head terms", "10–20 neutral prompts"), and edge cases in every section — with no padding and no explanation of concepts Claude already knows. Not 4: there is no passage of over-explanation to trim. | 5 / 5 |
Actionability | Concrete guidance is pervasive: named MCP tools with settings ("call `explore_prompt` with its most representative prompt, the default ChatGPT model and `highlightBrand` set to the brand"), a copy-paste JSON payload for `update_project_context`, an exact merge rule ("lowercase, trimmed, collapsed whitespace"), and a full report template. It misses 5 because the argument shape of the primary `research_ai_visibility_prompts` call is never shown. | 4 / 5 |
Workflow Clarity | Five numbered steps in clear sequence with explicit checkpoints and feedback loops: "summarize each result before the next call", "If a term returns nothing, retry once with a shorter or broader form before dropping it", "ask once whether to spend a small amount of credit", and "If the call fails, such as no paid plan, skip it and say so". Not 4: error recovery and spend-confirmation checkpoints are explicit. | 5 / 5 |
Progressive Disclosure | A single well-sectioned file with clear headers and no nested or buried references, so navigation is easy. It scores 4 rather than 5 because at ~95 lines it exceeds the under-50-line simple-skill case, and content that could live in one-level-deep references (tool semantics, report template) is inlined. | 4 / 5 |
Total | 18 / 20 Passed |