Content
62%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
The skill is highly actionable with excellent workflow clarity, providing concrete JSON examples, specific MCP tool names, and a well-sequenced multi-step process with validation. However, it is significantly over-verbose — the enabled=False/targeting warning is repeated at least 4 times (top warning, Step 5, checklist, What NOT to Do), and several sections explain things Claude can infer. The content would benefit greatly from deduplication and trimming.
Suggestions
Consolidate the enabled=False/targeting warning to a single prominent location (Step 5) instead of repeating it in the top warning block, Step 5 body, the verbatim checklist, and the 'What NOT to Do' section.
Remove or drastically shorten the top warning block — the same information is covered more actionably in Step 5.
Trim the 'What NOT to Do' section to only items not already covered in the workflow steps, or remove it entirely since the workflow already provides clear guidance.
Consider extracting the targeting shortcut (curl/CLI commands) into a separate referenced file to reduce the main skill's length.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill is extremely verbose at ~250+ lines. The giant warning block at the top, repeated explanations of the targeting/enabled=False issue (mentioned at least 4 times), the full verbatim checklist to print, and explanations of concepts Claude can infer (like what agent vs completion mode means structurally) all waste significant tokens. The 'What NOT to Do' section repeats guidance already given in the workflow steps. | 1 / 3 |
Actionability | The skill provides fully executable JSON payloads for both agent and completion modes, concrete curl commands for the targeting shortcut, specific MCP tool names with exact parameter formats, and clear examples of modelConfigKey values. The guidance is copy-paste ready throughout. | 3 / 3 |
Workflow Clarity | The workflow is clearly sequenced (Steps 1-5) with explicit validation checkpoints (Step 4 verification, Step 5 targeting confirmation). It includes feedback loops for error cases (NO MODEL fix, jq shape inspection), a decision matrix for mode selection, and explicitly calls out the most common failure mode with a concrete resolution path. | 3 / 3 |
Progressive Disclosure | The skill references related skills (tools, configs-variations, configs-update, configs-targeting) but all content is inline in one monolithic file. The agent vs completion decision matrix, the modelConfigKey reference, the edge cases table, and the targeting shortcut commands could all be split into referenced files. However, there are no bundle files to split into, and the references to sibling skills are clearly signaled. | 2 / 3 |
Total | 9 / 12 Passed |