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: a clearly sequenced five-step workflow with state branching, spend-approval gates, error recovery, and evidence-verification checkpoints, plus fully executable guidance naming exact tools, parameters, and report structure. The only weaknesses are mild — a few sentences could be tightened, and some late-phase detail could be split into one-level reference files to slim the main file.
Suggestions
Move the report-format specification in 'What to return' (section list, table columns, example chat sentence) into a one-level reference file such as references/report-format.md, keeping a short pointer in SKILL.md so the main file stays a process overview.
Extract the five tracker-state branches in workflow step 1 into a short decision table or reference file, leaving the common path (latest completed run) inline.
Tighten the longest sentences, such as project-context item 3 on brand-name confirmation, into short imperative bullets to save tokens without losing meaning.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense, directive prose with no explanations of concepts Claude already knows (e.g., the robots.txt bot list explains domain-specific live-answer vs training semantics Claude cannot be assumed to know). A few long, winding sentences — project-context item 3 on brand-name confirmation and the no-tracker branch in step 1 — could be tightened, matching anchor 4 (efficient, minor trim opportunities) rather than the every-token-earns-its-place anchor 5. | 4 / 5 |
Actionability | Quotes: "Pass `limit: 50`, and when `totalCount` is larger than the rows returned, follow `nextCursor`", "run one check with the approved `runNowCostUsd` as `maxCostUsd`", "Read two or three of the gap answers with `get_ai_visibility_answer`". Exact tools, parameters, counts, a fully specified report structure with section names and table columns, and an example chat sentence make the guidance fully executable for an instruction-only skill, matching anchor 5. | 5 / 5 |
Workflow Clarity | Five sequenced workflow steps with explicit branches for every tracker state, approval gates before any spend ("get approval once"), error-recovery loops ("A failed or partial run still supplies whatever answers it completed"; "If it returns an error, such as no paid plan, continue without it and say so"), verification checkpoints ("Read it before claiming so"), and a completeness rule ("Every candidate ends as a recommendation or a row..."). Explicit validation and feedback loops match anchor 5 rather than the minor-gaps anchor 4. | 5 / 5 |
Progressive Disclosure | The skill is a single well-sectioned file (Goal, Project context, Tools, five Workflow steps, What to return, Guardrails) with no bundle files, and all content is needed during a run, so nothing is buried or nested. However, at roughly 104 lines, terminal-phase detail such as the report-format specification and the five tracker-state branches could live in one-level reference files to keep the overview leaner, matching anchor 4's 'good structure; minor organization gaps' rather than anchor 5, whose simple-skill exception (under 50 lines) does not apply. | 4 / 5 |
Total | 18 / 20 Passed |