Proficiency Levels

Your Skill score maps to one of five proficiency tiers. Skill is on a 0–100 scale and comes from five inputs (SSE, PES, IE, CRR, FC) — see PRISM Scores → Skill for the formula.

Range	Tier
90 – 100	Elite
70 – 89	Expert
50 – 69	Proficient
30 – 49	Practitioner
0 – 29	Novice

On the equivalent 0–10 grade scale, B (7.0–7.9) is baseline — the team average sits there. Expert on Skill corresponds to solid-B territory on the grade scale.

Reading each tier

The descriptions below use the five Skill inputs as the frame: SSE (outcome efficiency), PES (prompt quality), IE (iteration turns per hour), CRR (context reset cadence), FC (focus stretch length).

Novice (0–29)

• SSE — outcomes far below your own baseline: many tokens, many turns, long sub-sessions for small results
• PES — prompts lack file paths, function names, or expected behavior; often bundle multiple asks
• IE — many turns per active hour; retry storms when things fail
• CRR — rarely /compact or /clear, or resets randomly
• FC — short, fragmented stretches

Practitioner (30–49)

• SSE — sometimes matches baseline, often below on bundled or under-specified asks
• PES — some specificity, still frequently vague
• IE — limited recovery strategy; tends to retry rather than add constraints
• CRR — starting to use /compact but inconsistent
• FC — single focused stretches occasionally, broken up by context switches

Proficient (50–69)

• SSE — meets baseline on routine work; dips on unfamiliar problems
• PES — prompts include file paths and function names; mostly single-task
• IE — recovers from errors by adding constraints, not repeating prompts
• CRR — resets at natural breakpoints between tasks
• FC — sustained stretches on focused work, usually on planned tasks

Expert (70–89) — team-average / baseline

• SSE — consistently at or above baseline; sub-sessions reach goals with minimum-path efficiency
• PES — specific prompts with constraints and expected behavior; trusts CLAUDE.md instead of re-explaining
• IE — low turns-per-hour; prompts converge toward solutions
• CRR — reset cadence lands on the healthy-middle band
• FC — long deep-work stretches are the norm

Elite (90–100)

• SSE — regularly beats baseline by a wide margin; minimal rework
• PES — prompts read like well-scoped tickets; information density is high and ambiguity cost is low
• IE — near-minimum turns; solutions land in one or two turns even on non-trivial work
• CRR — context management is second nature, never a drag on flow
• FC — consistently long focused stretches; rarely interrupted

What lifts your tier the most

The highest-leverage moves at each step — keyed to the input that moves Skill the most for typical developers at that level:

From	Focus on	Why it works
Novice → Practitioner	Add file paths and function names to prompts	Lifts PES directly and almost always lifts SSE through shorter sub-sessions
Practitioner → Proficient	One task per prompt; paste the error on failure	Splits bundled asks (PES Turn Economy) and cuts retry storms (IE Recovery)
Proficient → Expert	Use plan mode and CLAUDE.md; state expected behavior	Raises PES Context Leverage and reduces IE turns-per-hour
Expert → Elite	Delegate with subagents; master context resets	Optimizes FC and CRR — the last two inputs most developers leave on the table

The biggest gains come early. Use /prism:score to see your weakest input and the next concrete change to make.