Two capabilities

Prism has three goals on the user side: ship more finished work, use tokens well, and get better at prompting. The product groups its capabilities into two halves that serve those goals — under one headline number, the Prompt Score (average prompt grade, 0–100), with a supporting Completion rate beneath it.

Goal	Metric	What moves it
Get better at prompting	Prompt Score (APG) — headline	Prompt Advisor (coaching) + Usage Intelligence (habit tracking)
Ship more finished work	Completion rate · Speed (CSPW)	Prompt Advisor (better prompts land their goal)
Use tokens well	Token Usage (Crush Weight, TPT)	Usage Intelligence (cost visibility) + Prompt Advisor (shorter paths)

See Prompt Score for the headline score and Algorithm Overview for how a sub-session gets graded as crushed.

Prompt Advisor

Audience: individual users.

Helps you write better prompts and improve your AI-collaboration habits — live on the dashboard and after the fact. The coaching lives in the dashboard, not the terminal; the plugin’s job is to capture telemetry and prompt text.

Feature	Description
Realtime view	The dashboard realtime view streams each prompt as you send it, scored turn-by-turn, with a per-sub-session Task Advisor (weakest rubric check plus an example rewrite) — see Realtime
Live session snapshot	/prism:realtime prints the current session’s live grade, cost, and turn count in the terminal
Session advisor cards	LLM-authored coaching cards per sub-session, tied to the weakest prompt habit — see Advisor & Summaries
Session title & summary	LLM-generated title and 1–2 sentence recap on every session, editable from the dashboard
Skill profile	Skill Index with per-pillar trends — see Skill
Report	Chapter-based personal review with strengths, weaknesses, worst prompts, and heatmaps — see Report

Usage Intelligence

Audience: users first; team leads for aggregate views.

Turns raw telemetry into outcome, cost, and habit signals you can act on.

Feature	Description
Prompt Score	Headline score — Average Prompt Grade (APG) over time — see Prompt Score
Completion rate	Crushed-sub-session rate over time (supporting outcome metric)
Speed	Crushed sub-sessions per window with TTC (Time To Crush) — see Speed
Token Usage	Crush Weight, TET, TPT, and token mix — see Token Usage
Prompt Grade	Per-prompt grading with a per-session breakdown — see Prompt Grade
Trivia	Sub-sessions filtered out by the substance floor — see Trivia
Integrity	Anti-gaming flags raised against your sessions — see Integrity
Sub-session explorer	Browse and drill into any closed sub-session, turn-by-turn

How the two work together

User writes prompt
    → Telemetry + prompt text captured (Ingest → API)
    → Session closes on /clear or topic shift
    → API grades the prompt and the outcome judge marks it crushed / not crushed
    → Usage Intelligence updates Prompt Score, Completion rate, Speed, Token Usage
    → Prompt Advisor surfaces coaching on the dashboard (realtime Task Advisor, advisor cards, Report)
    → User improves → cycle repeats

Better prompts (Advisor) raise your Prompt Score and make sub-sessions more likely to land their goal, which raises the Completion rate and pulls Crush Weight down (fewer tokens per finished sub-session). The two aren’t separate products — they’re two views of the same loop.