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 is exceptionally actionable with executable SQL templates and a clearly sequenced, validated workflow. Its main weaknesses are conciseness (some repetition of the core discriminator) and progressive disclosure (a large monolithic file with no external reference split).
Suggestions
Extract the optimization-lane SQL probes and footguns into a references file (e.g., OPTIMIZATION.md) referenced one level deep, leaving SKILL.md as an overview.
Dedupe the 'configured-to-sync vs actually-syncing' discriminator, which appears in the intro, Explore, and Disqualifiers — state it once and reference it.
Move the MCP tools catalog into a separate references file to reduce SKILL.md token load.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Highly dense with load-bearing specifics and no padding of concepts Claude already knows, but the core discriminator and several optimization-lane footguns are restated across the intro, Explore, and Disqualifiers sections, so it could be tightened. | 3 / 5 |
Actionability | Fully executable — complete copy-paste SQL templates with exact filters, concrete tool calls with arguments, exact scratchpad key formats, priority mappings, and specific error-class to root-cause mappings covering the common cases. | 5 / 5 |
Workflow Clarity | Clear sequenced run workflow (orient → profile → explore → save memory → decide → close out) with explicit validation checkpoints, edit-vs-author decision rails, 7-day/≥3-active-day recheck loops, and explicit stop criteria. | 5 / 5 |
Progressive Disclosure | No bundle files exist and the file has good internal section headers, but the optimization-lane SQL probes, footguns, and MCP tools reference are inlined in one monolithic SKILL.md rather than split into one-level-deep reference files. | 3 / 5 |
Total | 16 / 20 Passed |