Content
50%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.
The body is a well-organized but monolithic API catalog: every endpoint has real curl examples, yet a meaningful fraction are syntactically broken or use parameter keys inconsistent with their own documented signatures. The 300+ lines of reference belong in a separate file, and there is no decision guidance for picking among the many overlapping endpoints. Fixing the malformed examples and splitting out the reference would raise both actionability and progressive disclosure.
Suggestions
Fix the broken curl examples: wrap searchParams inside the -d JSON body with correct quoting for investor/people/company/job search, and align example keys with documented parameters (profileIdentifier/companyIdentifier instead of linkedin_url/domain, contentId instead of identifier for post-comments, type/value for live company fetch).
Move the per-endpoint parameter reference into a references/ file (e.g., references/endpoints.md) and keep SKILL.md as an overview with setup, cost notes, and one worked example, removing the duplicated Capabilities bullets.
Add endpoint-selection guidance (when to use natural-language search vs. people/company search filters vs. kitchen-sink) and a basic response-check step (e.g., verify HTTP status/record count before paginating).
Complete the truncated capability bullets ("Takes free-form text (e") and trim the repeated curl header boilerplate by noting auth once.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is mostly efficient reference material, but the 17 endpoint sections repeat identical curl boilerplate (auth headers, proxy invocation) verbatim, the Capabilities bullet list duplicates the Usage section headers nearly word-for-word, and two bullets are literally truncated mid-sentence ("Takes free-form text (e"). This matches "mostly efficient but includes some unnecessary explanation or could be tightened"; not a 4 because the duplication and truncation artifacts are more than minor. | 3 / 5 |
Actionability | Concrete curl commands exist for every endpoint, but several are broken: the investor/people/company/job search examples are invalid shell/JSON ('-d '{"api":"fiber","path":"..."}' "searchParams": {...}'' — the searchParams sits outside the JSON body with mismatched quotes), and multiple examples use keys the documented parameters don't define (linkedin_url/domain instead of profileIdentifier/companyIdentifier; identifier instead of contentId for post-comments, and instead of type/value for live company fetch). This is "some concrete guidance but incomplete; missing key details"; not a 4 because multiple examples fail as written. | 3 / 5 |
Workflow Clarity | A setup sequence exists (read credentials, export env vars, check for missing credentials file with a fallback instruction), but there is no guidance on choosing among the 17 overlapping endpoints (only one aside recommending kitchen-sink over email lookup) and no validation of responses or error-handling steps. This matches "steps listed but validation gaps; checkpoints missing or implicit"; not a 2 because the setup steps are well defined and executable. | 3 / 5 |
Progressive Disclosure | The file has clear section headers and a "Discover More" pointer to live endpoint details, but ~300 lines of per-endpoint API reference (parameters and examples for 17 endpoints) are inlined in SKILL.md with no bundle files present — content that clearly belongs in a references/ file. This matches "some structure but could be better organized; content that should be separate is inline"; not a 4 because the bulk of the skill is inlined reference material rather than an overview. | 3 / 5 |
Total | 12 / 20 Passed |