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.
The body is a tight, actionable workflow: a step-numbered checklist, a concrete actor selection table, exact commands for schema fetching and three output modes, and an error-handling table with recovery actions. The main weaknesses are the absence of a concrete example actor input and an explicit output-validation step before reporting results.
Suggestions
Add one concrete example of a filled --input value (e.g., '{"searchStrings": ["coffee shops in Austin"], "maxCrawledPlaces": 20}') so JSON_INPUT is not left to synthesis.
Insert an explicit validation checkpoint in Step 4/5, e.g. verify the output file exists and is non-empty (or check the returned item count) before reporting success.
Trim low-value filler such as "(No need to check it upfront)" and the vague "Based on character of use case", and consider moving the 17-actor table to a references file if the skill grows.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean: a data-dense actor selection table, exact bash commands, and a terse error table, with only minor trimmable padding such as "(No need to check it upfront)", "Number of results: Based on character of use case", and the copy-this-checklist block. Not 5 because those fluff fragments exist, but well above the 'could be tightened' level 3. | 4 / 5 |
Actionability | Provides concrete, mostly copy-paste-ready commands (mcpc tools-call invocation, node run_actor.js with full flags per output format, error-recovery responses). Not 5 because JSON_INPUT is never exemplified — no sample actor input appears anywhere — and --timeout is referenced only in error handling without appearing in a usage example. | 4 / 5 |
Workflow Clarity | Clear five-step sequence with a progress checklist, a schema-fetch step that gates input construction, error-recovery guidance, and a results-summary step. Not 5 because there is no explicit output validation checkpoint (e.g., verify the CSV/JSON is non-empty or check lead count before reporting); the batch-operation cap at 3 is not triggered since error feedback loops and a verification-oriented Step 5 are present. | 4 / 5 |
Progressive Disclosure | Well-organized single-file structure with clear sections, and fetching actor schemas dynamically via mcpc elegantly avoids inlining per-actor documentation. Minor gaps: the 17-row actor table is borderline reference material that could live in a separate file, and the referenced ${CLAUDE_PLUGIN_ROOT}/reference/scripts/run_actor.js path cannot be verified within the skill bundle. | 4 / 5 |
Total | 16 / 20 Passed |