Content
85%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.
A highly actionable, well-structured MCP command reference: exact tool names, parameters, cost-guardrail validation loops, and clean one-level-deep bundle references (both referenced files exist and hold the promised content). The main weakness is token efficiency — a caveat repeated three times, an approval flow duplicated between the body and cost-tiers.md, and date-stamped version pins outside any deprecated section.
Suggestions
State the 'pinned MCP server has no Google Images tool' caveat once (in the serp-images section) and reference it from the quick-reference table and the seo-images cross-skill bullet instead of repeating the full explanation.
Keep the cost-approval flow in only one place (references/cost-tiers.md) and have the Cost Guardrails section say just 'run the check script before every call; see references/cost-tiers.md for the approval flow and outcomes', removing the duplicated branch/log steps and user commands.
Move the version pin and verification dates (dataforseo-mcp-server@2.8.10, 2026-09-23) into a short 'Version notes' or deprecated-style section so the time-sensitive stamps don't sit in the main operational text.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is a dense, competent command catalog with no basic-concept padding, but it could be tightened: the 'no Google Images tool on the pinned MCP server' caveat is repeated three times (quick-reference table, the serp-images section, and the seo-images cross-skill bullet), and the cost-approval flow (check/approved/needs_approval/blocked/log) is fully duplicated in both the body and references/cost-tiers.md. Time-sensitive version/date stamps ("dataforseo-mcp-server@2.8.10", "checked against the package, 2026-09-23") appear outside any deprecated/old-patterns section, which the rubric explicitly penalizes, holding this below the level-4 'minor instances that could be trimmed' anchor. | 3 / 5 |
Actionability | Every command specifies exact MCP tool names (e.g., `serp_organic_live_advanced`, `ai_opt_llm_ment_search`), concrete default parameters (location_code=2840, language_code=en, depth=100), expected outputs, and copy-paste-ready shell commands for cost guarding ("${CLAUDE_PLUGIN_ROOT}/scripts/claude-seo" run dataforseo_costs.py check <endpoint>). As an instruction-only MCP skill its guidance is fully executable and covers the common cases, matching the level-5 anchor. | 5 / 5 |
Workflow Clarity | Multi-step processes are explicitly sequenced with validation checkpoints: a mandatory pre-call cost check with three-way branching (approved → proceed, needs_approval → confirm with user, blocked → do not proceed), post-call cost logging, a numbered 4-step workflow for ai-mentions, and an error-handling section with recovery guidance (rate limit → wait and retry; no results → report and broaden query). The explicit pre-call validation and error feedback loops match the level-5 anchor; the batch/destructive cap does not apply since validation is present. | 5 / 5 |
Progressive Disclosure | The body is a well-navigable overview (quick-reference table up front, grouped per-command sections) that appropriately offloads the pricing table to references/cost-tiers.md and the bulk utility tool list to references/tool-catalog.md — both real files, clearly signaled ("Load references/cost-tiers.md for the full pricing table"), exactly one level deep with no nested references. This matches the level-5 anchor: clear overview with well-signaled one-level-deep references and appropriately split content. | 5 / 5 |
Total | 18 / 20 Passed |