Content
100%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 a tight, opinionated workflow: concrete sweep tactics, a hard reel-ability gate, a defined shortlist format, an explicit user checkpoint, and a precise save schema. It is concise, actionable, well-sequenced with validation, and cleanly structured with only one signaled external reference.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Lean and assumes Claude's competence throughout — no definitions of basic concepts, no library tutorials; every line states a standard, a test, or a concrete artifact, with lines like 'Recency is a fact about the fact' encoding a rule rather than padding. | 5 / 5 |
Actionability | Provides fully executable guidance: named search platforms, a query count range (six to ten), a five-part reel-ability test, a four-column shortlist table, and a copy-paste-ready JSON record with all fields populated for backlog/ideas.json. | 5 / 5 |
Workflow Clarity | A clear four-step sequence (Sweep → Judge → Shortlist → Save) with an explicit user-pick gate ('Never add to the backlog without an explicit pick') and an embedded validation checkpoint (the three-of-five reel-ability test that rejects candidates before listing). | 5 / 5 |
Progressive Disclosure | Self-contained, well-sectioned body with a single clearly signaled one-level-deep external reference ('Read .claude/skills/reel-script/voice.md first') and an explicit output file (backlog/ideas.json); no nested references and no inlined content that belongs elsewhere. | 5 / 5 |
Total | 20 / 20 Passed |