Content
52%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 well-organized with concrete thresholds, quantified budgets, sequenced workflows, and useful rationalization/red-flag/verification sections, but it is held back by structural duplication and, critically, an Advanced Documentation section whose six module references do not match the actual bundle (references/reference.md and references/examples.md are never linked). Fixing the reference paths and providing templates for the handoff package and session snapshot would lift it substantially.
Suggestions
Replace the six nonexistent modules/*.md pointers with links to the actual bundle files (references/reference.md and references/examples.md), or add the module files so the Advanced Documentation section resolves.
Deduplicate the overlapping sections — merge the two "When to Use" lists and the "Quick Reference" capabilities with the "Implementation Guide" features into a single section, and drop the abbreviated "Core Patterns" summary that is fully restated in "Core Patterns Detail".
Add concrete templates or examples for the handoff package structure (Pattern 4) and session state snapshot (Pattern 3) so the guidance is copy-paste executable rather than descriptive.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is mostly list-form and avoids explaining concepts Claude already knows, but it is padded with duplicated sections: two near-identical "When to Use" lists ("Managing large codebases exceeding 150K token limits" vs "Context window approaching limits exceeding 150K tokens"), overlapping capability/feature lists in "Quick Reference" and "Implementation Guide", and "Core Patterns" summarized in the Implementation Guide then re-expanded in "Core Patterns Detail". This fits anchor 3 (mostly efficient but could be tightened) rather than 2, since the individual sections are terse and the duplication is structural rather than fluffy prose. | 3 / 5 |
Actionability | There is genuinely concrete guidance — specific thresholds (warning at 150K, critical at 180K, clear at 85%), quantified budget breakdowns, and an itemized handoff package field list — but everything is described rather than templated: no copy-paste command sequences, no handoff package format example, no session-state snapshot template. As an instruction-only skill this lands at anchor 3 (some concrete guidance but incomplete) rather than 4, which would require mostly executable, specific instruction blocks. | 3 / 5 |
Workflow Clarity | Multi-step processes are sequenced ("Session Initialization", "SPEC-First Workflow: Execute /moai:1-plan, then mandatory /clear to save 45-50K tokens, then /moai:2-run...") and validation is present (handoff validation checks, a Verification checklist with checkboxes, and monitoring thresholds with defined actions at 60/75/85%). This matches anchor 4 (clear sequence, most checkpoints present) rather than 5 because the workflow sequences are compressed run-on paragraphs instead of numbered steps with explicit error-recovery loops. | 4 / 5 |
Progressive Disclosure | The "Advanced Documentation" section points to six module files (modules/token-budget-allocation.md, modules/session-state-management.md, modules/context-optimization.md, modules/handoff-protocols.md, modules/memory-mcp-optimization.md, modules/reference.md), but none of these exist in the bundle — the actual files are references/reference.md and references/examples.md, neither of which is referenced from the body. Every navigation pointer is broken, and substantial detail (the full pattern descriptions, budget tables) is inlined that belongs in those separate files. This matches anchor 2 (references buried/broken, content that should be separate is inline) rather than 3, because the structure exists on paper but resolves to nothing on disk. | 2 / 5 |
Total | 12 / 20 Passed |