Content
65%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
A highly actionable, well-sequenced research skill with concrete tool calls and a usable report template. Weaker on workflow validation checkpoints (lives outside the flow) and progressive disclosure (monolithic, no external references).
Suggestions
Embed a verification checkpoint inside the workflow (e.g. Step 5.5: cross-reference claims and flag any sub-question with insufficient sources before writing the report) rather than relegating it to Quality Rules.
Move the full report markdown template and/or the example prompts into a referenced file (e.g. REPORT_TEMPLATE.md) so SKILL.md stays a lean overview with one-level-deep navigation.
Remove or trim the 'Examples' section since its prompts overlap with 'When to Activate', recovering tokens without losing trigger coverage.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is largely action-dense and avoids explaining known concepts, but the trailing 'Examples' section duplicates 'When to Activate' and the full inline report template is verbose, so it could be tightened. | 2 / 3 |
Actionability | Concrete, parameterized MCP calls (firecrawl_search limit:8, web_search_advanced_exa with startPublishedDate), specific numeric targets (15–30 sources, 2–3 keyword variations), and a copy-ready report template give fully executable guidance. | 3 / 3 |
Workflow Clarity | The six steps are clearly sequenced, but verification/gap-checking guidance sits in a separate 'Quality Rules' section rather than as embedded checkpoints in this batch research workflow, so validation is implicit. | 2 / 3 |
Progressive Disclosure | At ~153 lines with no bundle files, the skill is a well-sectioned but monolithic document; the report template and example prompts are inline content that could be split into one-level references. | 2 / 3 |
Total | 9 / 12 Passed |