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.
A solid, action-oriented skill body: executable commands, a clear sequenced workflow with a progress checklist, and genuine error-recovery guidance. The main gaps are minor — unexemplified input JSON, an unverifiable script reference, triplicated command blocks, and no explicit output-verification step for batch runs.
Suggestions
Consolidate the three near-identical run commands into one block with documented optional flags (--output, --format) to save tokens.
Add one concrete --input example (e.g. a small JSON payload for compass/crawler-google-places) so the JSON_INPUT placeholder is unambiguous.
Add an explicit verification step in Step 5: confirm the output file exists and inspect a sample of rows before summarizing findings.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean — an actor-selection table, copy-paste commands, and terse error handling — with no explanations of concepts Claude already knows. Minor trimming is possible: the three near-duplicate run-command blocks (Quick/CSV/JSON) differ only in flags and could be consolidated, and the 'Best For' column largely echoes the 'User Need' column. This fits anchor 4 ('efficient; minor instances that could be trimmed') rather than 5. | 4 / 5 |
Actionability | Commands are concrete and executable with real flags (mcpc tools-call with auth header, node --env-file with --actor/--input/--output/--format), and Step 2's dynamic schema fetch is a smart executable pattern. It stops short of anchor 5 because 'JSON_INPUT' is never exemplified with a real payload, and the referenced script '${CLAUDE_PLUGIN_ROOT}/reference/scripts/run_actor.js' is not present in the skill bundle. | 4 / 5 |
Workflow Clarity | A clear 5-step sequence with a copyable progress checklist, and the Error Handling section provides explicit failure-to-remedy feedback loops (e.g. 'Run FAILED' - check Apify console link; 'Timeout' - reduce input or increase --timeout). Minor validation gap keeps it at anchor 4: for batch extraction runs there is no explicit output-verification step before reporting (e.g. confirm the file exists and inspect sample rows). | 4 / 5 |
Progressive Disclosure | Well-organized single-file structure (Prerequisites, 5-step Workflow, Error Handling) with the 16-row actor table appropriately inline as core selection logic. The single script reference is one level deep and clearly signaled, but it points to a plugin-root path with no scripts/ directory in the bundle, so navigation cannot be verified in-bundle — matching anchor 4's 'minor organization gaps' rather than 5. | 4 / 5 |
Total | 16 / 20 Passed |