Execution leverage
Turns ambiguity into useful questions, explicit decisions, and an actionable handoff.
Static rubric v5
How Scouter Score works.
A 10,000-point static assessment of how effectively a skill helps an agent understand, execute, and finish work. Popularity never changes the score.
25%
Procedural quality
Provides a concrete, coherent, and adaptable method the agent can follow.
Reliability
Validates inputs and outcomes, defines completion, and handles failure safely.
Outcome value
Produces a result that materially improves the task when the skill succeeds.
Applicability
Rewards portable guidance; framework-specific skills score lower and single-vendor APIs lower still.
Efficiency
Uses steps, context, and tokens in proportion to the work.
Highest-weight signal
What execution-ready looks like
- Defines the intended outcome
- Surfaces constraints and assumptions
- Clarifies priorities and tradeoffs
- Asks questions that change what happens next
- Establishes acceptance criteria
- Produces an actionable handoff
- Knows when to stop or escalate
Scoring boundary
Quality is not popularity
Recorded installs, GitHub stars, forks, and repository activity are shown as ecosystem evidence. They never increase Scouter Score.
Platform specificity
A focused skill can still be excellent, but its overall usefulness is narrower. Portable guidance typically scores 90–100 for Applicability, framework-specific guidance 55–70, and a single vendor product or API 25–50.