Content
63%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 thorough, well-sequenced operational skill with genuinely executable API specs and strong honesty/confirmation guardrails around paid scraping. Its weaknesses are length and monolithic structure: duplicated trigger lists and a full report template inlined in SKILL.md instead of split into reference files, and two platform phases that defer endpoint details to an external doc.
Suggestions
Split the ~150-line Phase 8 report template and the per-platform scraping details into reference files (e.g., references/report-template.md, references/google-ads.md) and link to them from a lean overview, moving progressive disclosure from inline monolith to one-level-deep references.
Merge the 'When to Use' and 'Trigger Phrases' sections, which repeat nearly the same list of prompts, and drop the intro sentence that restates the frontmatter description verbatim.
Give the TikTok and LinkedIn phases the same concrete endpoint call specs (path, query parameters) that the Meta and Google phases have, so all four channels are equally executable without consulting an external reference.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is mostly efficient, domain-specific process content, but the 'When to Use' and 'Trigger Phrases' sections duplicate nearly the same list, and the ~150-line Phase 8 report template is inlined in full — it could be tightened or externalized. Not anchor 2 because there is little explanation of concepts Claude already knows, and not anchor 4 because the duplication and inline template are noticeable slack. | 3 / 5 |
Actionability | Mostly executable guidance: exact YAML provider call specs with paths and query parameters for Meta and Google, a copy-paste JSON tool invocation with validated limits, concrete per-ad collection checklists, and structured analysis tables. Minor gaps keep it below 5 — Phases 2 and 4 only say to "resolve the current TikTok/LinkedIn Ad Library search and ad-detail operations from the official provider reference" without giving the endpoints, unlike the Meta phase. | 4 / 5 |
Workflow Clarity | A clear Phase 0–8 sequence with most checkpoints present: user confirmation of advertiser identity, explicit spend/request limits, polling a stable source_id with job_get, and honest reporting of queued/running/partial/empty/failed states plus documented fallbacks. Not 5 because validation outside the Google import path is implicit (e.g., no verify step that scraped ad records are complete before analysis). | 4 / 5 |
Progressive Disclosure | Section structure is clear and well-organized, but there are no bundle files at all and roughly 450 lines of content inlined — the full report template and per-platform scraping details clearly belong in separate reference files. This matches the 3-anchor (structure present, but content that should be separate is inline); it is above 2 because navigation via headers is easy and nothing is buried. | 3 / 5 |
Total | 14 / 20 Passed |