Thoroughly Rate Review

Quick Index (Action-Routed)

Read First (All Actions)

1. Mission
2. Scope Boundary
3. Trigger Rule (Required)
4. Core Method

Action Modules (Read As Needed)

1. Building weighted rubric:

• Weighted Framework Standard
• Assessment Structure
• Scoring Anchors (0-5)

1. Evidence and confidence handling:

• Evidence Discipline

1. Mode-specific depth:

• Review Modes

1. Comparing multiple options:

• Comparison Mode (Optional)

Output

1. Output Contract (Required)
2. Anti-Patterns

Mission

Provide rigorous, explainable, and repeatable ratings instead of ad-hoc opinions.

Scope Boundary

This skill evaluates quality with weighted scoring.

Do not use this skill to choose governance mode/gates for execution planning. Use Skill Governance for risk-mode and gate policy decisions.

Trigger Rule (Required)

Use this skill whenever user intent is to evaluate quality, including synonyms such as:

1. review
2. rate
3. score
4. assess
5. evaluate
6. grade
7. benchmark
8. compare quality
9. audit quality

If intent is ambiguous:

1. assume evaluation intent when user asks “how good,” “how strong,” or “how would you rate”
2. proceed with the weighted framework

Core Method

1. define weighted categories totaling 100%
2. define multiple assessment checks per category
3. score each check on a fixed scale
4. compute category and weighted totals
5. provide final score with strengths, gaps, and upgrade path

Weighted Framework Standard

Default unless domain requires custom weights:

1. Coverage and Completeness: 14%
2. Correctness and Internal Consistency: 18%
3. Practical Utility and End-User Value: 14%
4. Safety, Risk Control, and Reliability: 14%
5. Enforceability and Operationalization: 14%
6. Clarity and Documentation Quality: 10%
7. Efficiency and Maintainability: 6%
8. Integration and Cohesiveness: 10%

All weights must sum to exactly 100%.

Mandatory rule:

1. for multi-skill systems/frameworks, Integration and Cohesiveness is required and cannot be omitted

Assessment Structure

For each category:

1. define 3-6 concrete checks
2. score each check 0-5
3. compute category raw score:

• (sum(check scores) / max possible) * 100

1. compute weighted contribution:

• category raw score * category weight

Final score:

1. sum all weighted contributions
2. round to one decimal place
3. present as X.Y / 100

Integration and Cohesiveness checks (minimum set):

1. cross-component/skill trigger alignment
2. conflict-resolution compatibility
3. sequencing compatibility across scenarios
4. policy/runtime enforcement consistency

Scoring Anchors (0-5)

1. 0: absent or broken
2. 1: major deficiencies
3. 2: partial and unreliable
4. 3: adequate baseline
5. 4: strong and consistent
6. 5: excellent and robust

Evidence Discipline

Every non-trivial score must include evidence:

1. direct references to files/behaviors/artifacts
2. brief rationale for each category score
3. explicit assumptions when evidence is incomplete

If evidence is missing:

1. lower confidence
2. cap category score to avoid over-claiming
3. state what would raise confidence

Output Contract (Required)

Return in this order:

1. framework used (weights + checks summary)
2. category-by-category table
3. final score
4. top strengths
5. top gaps
6. prioritized actions to improve score

Review Modes

1. quick:

• fewer checks per category, lower evidence depth

1. standard:

• full weighted model and concise evidence

1. deep:

• full weighted model + exhaustive evidence + scenario analysis

Default mode:

1. use standard
2. use deep if user asks “thorough,” “detailed,” or similar

Comparison Mode (Optional)

When comparing options:

1. use same weighted framework for all options
2. score each option independently
3. include normalized side-by-side table
4. call out tradeoffs, not just winner

Anti-Patterns

1. giving a score without explicit framework
2. changing weights mid-review without explanation
3. rating from intuition without evidence
4. ignoring user context or constraints
5. presenting a single number without actionable next steps

Related Skills

• Thoughtful Approach: improves end-user oriented judgments in the scoring model.
• Skill Governance: provides risk-aware process context for evaluations.
• Doc Maintenance: ensures rating rationale is captured in durable docs when needed.
• Token Reduction: keeps scoring outputs concise while preserving evidence.