Content
76%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 skill body with concrete commands, realistic configuration, and a clear phase sequence. Its main weakness is the absence of validation and error-recovery steps for what is a batch scraping workflow — failed profiles, empty results, and API errors are unhandled — plus minor redundancy between the 'When to Use' and 'Trigger Phrases' sections.
Suggestions
Add validation and feedback loops after Phases 1 and 2: check each scrape result for failures, empty outputs, or auth/rate-limit errors, and specify a retry-or-skip policy before proceeding to clustering (this would lift workflow_clarity above the batch-operation cap of 3).
Merge the 'Trigger Phrases' section into 'When to Use' (or reference it) — the two sections duplicate the same trigger scenarios and cost tokens without adding coverage.
Move the ~60-line Phase 4 output template to a reference file (e.g. references/output-template.md) and keep a short summary inline, tightening the SKILL.md overview.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is efficient and assumes Claude's competence — no space is spent explaining what LinkedIn or topic clustering is — with lean commands, config, and templates. Minor trimmable padding remains: the 'Core principle' editorial line and the 'Trigger Phrases' section, which largely duplicates 'When to Use'. This matches the 4 anchor; it is not the 5 anchor because a few tokens do not earn their place, and not 3 because the padding is minor rather than a pattern. | 4 / 5 |
Actionability | Fully executable, copy-paste-ready guidance: concrete CLI invocations ('python3 skills/linkedin-profile-post-scraper/scripts/scrape_linkedin_posts.py --profiles ... --days ... --output json'), a complete config JSON with realistic example values, a full output template, a cron line, and a cost table. Placeholders like '<handle>' are parameterization, not pseudocode, matching the 5 anchor. | 5 / 5 |
Workflow Clarity | Phases 0-5 are clearly sequenced with post-scrape filters ('only include posts with reactions ≥ min_reactions'), but this is a batch operation — scraping 10+ profiles across two platforms — with no validation or feedback loops for failed scrapes, empty API responses, rate limits, or broken profile URLs. The rubric's batch-operation cap therefore limits this to 3 despite the good sequence; it stays above 2 because steps are well defined with few gaps. | 3 / 5 |
Progressive Disclosure | A single-file skill (no references/, scripts/, or assets/ directories exist) with well-organized sections, clear navigation, and no nested references; upstream-skill references are one level deep and clearly signaled. The ~60-line Phase 4 output template could arguably live in a separate reference file, which keeps this at the 4 anchor rather than 5. | 4 / 5 |
Total | 16 / 20 Passed |