Content
77%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 delivers an exceptionally actionable, well-sequenced debugging workflow with real tool calls, parameters, and a verification loop, backed by an executable snippets file. Its weaknesses are duplicated educational content and the orphaning of three of the four reference files, whose material is inlined rather than delegated.
Suggestions
Trim or cut the 'What is LCP and why it matters' intro and the subparts definition table — Claude already knows LCP thresholds, and the detailed breakdown lives in references/lcp-breakdown.md.
Replace the inlined Optimization Strategies section with a short prioritized summary that links to references/optimization-strategies.md for the full detail, mirroring how lcp-snippets.md is referenced.
Add clearly signaled links to the two unreferenced files (references/elements-and-size.md, references/lcp-breakdown.md) at the points in the workflow where their content applies, so all bundle files are discoverable.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body re-explains concepts Claude already knows ("LCP measures how quickly a page's main content becomes visible", the Good/Needs improvement/Poor thresholds, "LCP is a Core Web Vital that directly affects user experience and search ranking"), and the subparts table plus the ~35-line Optimization Strategies section duplicate material that exists in references/lcp-breakdown.md and references/optimization-strategies.md. This fits anchor 3 ("Mostly efficient but includes some unnecessary explanation or could be tightened"); the tool-specific workflow content itself is efficient, keeping it above anchor 2. | 3 / 5 |
Actionability | Guidance is fully executable: exact MCP tool calls with parameters ("performance_start_trace with pageId, reload: true and autoStop: true"), specific insight names (LCPBreakdown, DocumentLatency, RenderBlocking, LCPDiscovery), concrete network filters ("resourceTypes: [\"Image\", \"Font\"]"), specific fixes (preload with fetchpriority, font-display: swap, "Fast 3G" with cpuThrottlingRate: 4), and the referenced JavaScript snippets in references/lcp-snippets.md are real, complete, and copy-paste ready. Matches anchor 5. | 5 / 5 |
Workflow Clarity | The five-step debugging workflow is explicitly ordered ("Follow these steps in order. Each step builds on the previous one"), each step consumes the previous step's output ("Note the insight set IDs from the output — you'll need them"), conditional outcomes are handled ("If url is empty, the LCP element is text-based"), and the Verifying Fixes section provides an explicit feedback loop ("Re-run the trace... and compare the new subpart breakdown. The bottleneck should shrink"). Matches anchor 5's clear sequence with validation; the operations are non-destructive so no cap applies. | 5 / 5 |
Progressive Disclosure | Scored against the actual bundle: 4 reference files exist, but only references/lcp-snippets.md is ever linked from the body (clearly, twice). elements-and-size.md, lcp-breakdown.md, and optimization-strategies.md are never signaled, and their subject matter (subpart definitions, optimization strategies) is partially inlined in the body instead. This matches anchor 3 ("references present but not clearly signaled; content that should be separate is inline"); the body's clean section structure and the one well-signaled link keep it above anchor 2. | 3 / 5 |
Total | 16 / 20 Passed |