Content
75%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 well-structured, highly actionable skill document: every command and actor input is copy-paste executable against a real script, phases are clearly sequenced with fallback and error-recovery paths. The main costs are token weight from duplicated full-length output examples inlined in the body and implicit rather than explicit mid-workflow validation checkpoints.
Suggestions
Move one of the two full output examples (the JSON schema or the markdown summary — they overlap heavily) into a references/ file such as references/output-format.md, keeping only a short excerpt inline.
Add explicit validation checkpoints between phases, e.g. 'After Phase 1, confirm authority_score and organic keywords parsed non-null; if the scraper returned a login wall or empty data, switch to fallback mode.'
Trim the intro paragraph, which restates the frontmatter description almost word-for-word.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is mostly efficient — no explanations of basic concepts Claude already knows, and sections like Cost and Inputs are tight tables — but it could be trimmed: the full ~50-line sample JSON output and the separate ~40-line sample markdown summary largely duplicate each other as output-format specs, and the intro paragraph repeats the frontmatter description nearly verbatim. This fits 'mostly efficient but includes some unnecessary explanation or could be tightened' better than the level-4 'minor instances of over-explanation', since the two full output examples are a substantial trim opportunity. | 3 / 5 |
Actionability | Quick Start provides copy-paste-ready commands matching the actual script CLI (verified: --domain, --competitors, --keywords, --output, --skip-backlinks all exist in scripts/analyze_domain.py), each phase gives the exact Apify actor ID and concrete input JSON, and costs, dependencies, and the APIFY_API_TOKEN requirement are all explicit. Specific examples cover the common cases (basic run, competitor comparison, keyword checks, saved output), matching the fully-executable anchor. | 5 / 5 |
Workflow Clarity | The six phases are clearly sequenced with data dependencies stated (Phase 3 keyword sources in priority order, Phase 4 falling back to WebSearch when Semrush lacks per-page data) and there is an explicit error-recovery path ("Apify actors may break... fall back to the free seo-traffic-analyzer skill", plus a documented free-probe fallback when the token is missing). It falls short of the level-5 anchor because mid-workflow validation checkpoints are implicit rather than explicit — e.g., no instruction to verify the scraper actually returned parseable data before building the report — while the batch competitor runs are read-only, so the destructive/batch cap at 3 does not apply. | 4 / 5 |
Progressive Disclosure | Good structure with well-labeled sections (Quick Start, Inputs, Cost, Process phases, Output, Tips, Fallback, Dependencies) and a single referenced bundle file, scripts/analyze_domain.py, which exists and is correctly pathed. Scored against the actual bundle (scripts/ only, no references/): the main gap keeping it below 5 is that the long sample output blocks (~90 lines combined) are inlined in SKILL.md where a one-level-deep references/ file (e.g., output-schema.md) would keep the overview leaner; this fits 'good structure; most content appropriately placed; minor organization gaps'. | 4 / 5 |
Total | 16 / 20 Passed |