Content
52%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 excels at routing and orchestration logic — identifier detection, parallelization, person-to-company cascade, and output formatting are all clearly specified. However, a significant fraction of the curl examples contain malformed JSON or empty api/path placeholders that make them non-executable as written, the full-pipeline section duplicates the earlier reference sections almost entirely, and the skill is a monolithic single file where an endpoint reference and templates should be split out.
Suggestions
Fix the malformed curl examples: several close the -d JSON string early (Fiber name+company, Sixtyfour enrich-lead, find-phone, linkup, scrapegraph, exa) leaving raw JSON lines outside the quoted payload, and the Twitter, funding, and investors calls have empty '"api":"","path":""' placeholders — every example should be copy-paste executable.
Eliminate the ~120-line duplication in Section 4 (Full Pipeline Example), which re-inlines nearly every command from Sections 2 and 3; reference the earlier sections or keep only the differences (parameter variants, sequencing).
Split the skill into a lean SKILL.md overview plus one-level-deep reference files (e.g., references/person-endpoints.md, references/company-endpoints.md, references/output-templates.md) instead of the current ~540-line monolith.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The ~540-line body repeats the same curl boilerplate (auth headers, proxy URL) roughly 40 times and Section 4 ("Full Pipeline Example") re-inlines nearly every command already shown in Sections 2 and 3, so while there is little conceptual over-explanation, the duplication and padding could be cut substantially — fitting anchor 3 ("mostly efficient but could be tightened") rather than 4 given the scale of the repetition. | 3 / 5 |
Actionability | Concrete curl commands exist throughout, but many are malformed and not executable as written — e.g. the Fiber name+company call closes its -d string early and leaves raw JSON lines dangling ('"personName": {"fullName": "John Doe"}...}'), the same broken pattern appears in the Sixtyfour enrich-lead, find-phone, linkup, scrapegraph, and exa examples, and several calls contain empty placeholders ('{"api":"","path":""}') for the Twitter, funding, and investors endpoints — matching anchor 3's "pseudocode instead of executable code; missing key details" better than anchor 4's "concrete commands with minor gaps". | 3 / 5 |
Workflow Clarity | The sequence is clear and well-structured: an identifier routing table ("Email → Person + Company (extract domain)"), parallelization guidance, explicit cascade rule ("Person always cascades to company"), async polling instructions for Nyne, cross-referencing/conflict-flagging instructions, and a full worked example — anchor 4's "clear sequence with most checkpoints present". It is not 5 because there are no explicit error-recovery steps (e.g., what to do when an API call fails or returns empty) beyond conflict flagging. | 4 / 5 |
Progressive Disclosure | There are no bundle files at all (no references/, scripts/, or assets/ directories) and no external references from the body; the entire API endpoint reference, pipeline examples, and output templates are inlined in one ~540-line monolithic SKILL.md — matching anchor 2's "content that clearly belongs in separate files is inlined". It is not 3 because even a single well-signaled split (e.g., an endpoints reference or output-format template file) is absent. | 2 / 5 |
Total | 12 / 20 Passed |