This skill should be used when the user asks to "evaluate this output", "score this document", "assess quality", "rate this against criteria", "give me a structured evaluation", "check this before I share it", or when any output needs systematic quality assessment before use or publication. Provides a structured rubric-based evaluation framework applicable to any output type: documents, plans, code outputs, strategic analyses, research summaries, or pitches.
Structured, rubric-based evaluation for any output before it's used, shared, or published. Produces reproducible quality assessments with specific evidence and actionable improvements.
Apply evaluation when:
| Dimension | Weight | Criteria |
|---|---|---|
| Accuracy | 25% | Is the content factually correct? No hallucinations, outdated data, or false claims? |
| Completeness | 20% | Does it cover all necessary areas? What's missing? |
| Usefulness | 25% | Does it achieve its stated purpose? Better than alternatives? |
| Clarity | 15% | Is it readable, well-structured, unambiguous? |
| Freshness | 15% | Is the information current? Any stale references or outdated framing? |
| Score | Label | Criteria |
|---|---|---|
| 5 | Exceptional | Exceeds expectations; better than hand-crafted alternatives |
| 4 | Good | Minor gaps; clearly useful and fit-for-purpose |
| 3 | Adequate | Functional but notable room for improvement |
| 2 | Below average | Notable gaps or errors that undermine usefulness |
| 1 | Poor | Significant issues; needs substantial rework |
Before scoring, answer:
For each dimension, provide:
#### [Output Name] — Evaluation
| Dimension | Score | Evidence | Improvement |
|-----------|-------|----------|-------------|
| Accuracy | [1-5] | "[exact quote or specific example]" | [what to fix] |
| Completeness | [1-5] | "[what's missing or present]" | [what to add] |
| Usefulness | [1-5] | "[does it achieve its goal?]" | [how to improve] |
| Clarity | [1-5] | "[structural or language issues]" | [how to clarify] |
| Freshness | [1-5] | "[stale elements found]" | [what to update] |
| **Weighted Total** | **[X.X/5]** | | |
Weighted = (Accuracy × 0.25) + (Completeness × 0.20) + (Usefulness × 0.25) + (Clarity × 0.15) + (Freshness × 0.15)
| Score | Grade | Recommendation |
|---|---|---|
| 4.5–5.0 | A | Share/publish as-is |
| 3.5–4.4 | B | Minor polish recommended |
| 2.5–3.4 | C | Notable improvements needed before sharing |
| 1.5–2.4 | D | Significant rework required |
| < 1.5 | F | Start over or fundamentally rethink |
Always produce a ranked list of improvements:
### Improvements (Priority Order)
**MUST FIX** (blocks sharing):
1. [Specific issue] — [why it matters] — [how to fix]
**SHOULD FIX** (significant quality gain):
2. [Specific issue] — [why it matters] — [how to fix]
**NICE TO HAVE** (minor polish):
3. [Specific issue] — [why it matters] — [how to fix]
Add assessment of:
Add assessment of:
Add assessment of:
A good evaluation MUST: