FAQ
General
Section titled “General”What is Optra Prism?
Section titled “What is Optra Prism?”Optra Prism is an intelligence layer for AI-assisted coding. It plugs into Claude Code and gives you visibility into how you use AI — then helps you improve through prompt scoring, dashboard coaching, and usage analytics.
What data does it collect?
Section titled “What data does it collect?”- OTEL telemetry — logs and metrics (token counts, timing, model usage, tool calls)
- Prompt text — your actual prompts, used for prompt scoring and coaching
Prompt text is stored encrypted at rest and is scoped to your organization. See the privacy policy for details.
Does it slow down my coding?
Section titled “Does it slow down my coding?”The plugin hook is lightweight:
- Hook processing: sub-millisecond.
- Prompt capture to the ingest service: async, fire-and-forget.
- OTEL export: async, non-blocking.
The plugin makes no inline LLM calls, and it does not sit between Claude Code and Anthropic — your requests go straight there.
Does the plugin coach me inline in the terminal?
Section titled “Does the plugin coach me inline in the terminal?”No. There is no always-active advisor and no inline advice. The plugin captures telemetry; it does not score, coach, or rewrite prompts in your terminal. It can print a short realtime summary after a turn, but that is off unless you enable show_realtime_summary. Per-prompt scoring runs server-side, and coaching surfaces on the dashboard — the realtime view’s Task Advisor explains each sub-session’s grade and shows an example rewrite. For a detailed review, run /prism:report or read the chapter-based report on the dashboard.
Plugin
Section titled “Plugin”Do I need to restart Claude Code after setup?
Section titled “Do I need to restart Claude Code after setup?”Yes. The plugin writes OTEL_* variables into your settings.json on session start, and Claude Code reads those at process launch. Restart after running /prism:setup so they take effect.
Can Prism block my prompts?
Section titled “Can Prism block my prompts?”No. The submit hook always exits 0 and never modifies or delays your prompt. Everything the plugin emits is advisory.
What can I actually configure?
Section titled “What can I actually configure?”Two things: show_realtime_summary (off by default) and ingest_url (only if you run your own Prism). Everything else in ~/.prism/config.json is managed for you. See Configuration.
What if I don’t have an API key?
Section titled “What if I don’t have an API key?”Sign up for Prism at dashboard.optra-prism.com and finish onboarding — your key is issued automatically. Open the Connect panel in the dashboard header (Connect Claude Code); it shows the setup commands with your key already filled in, so there’s no key to create by hand and no separate keys page. New keys start with prism_* (the legacy gck_* prefix is still accepted).
How do I enable debug logging?
Section titled “How do I enable debug logging?”The debug log is always written to ${CLAUDE_PLUGIN_DATA}/debug.log (or ~/.prism/logs/debug.log if there’s no plugin data dir). To also mirror it to your terminal (stderr), set "debug": true in ~/.prism/config.json. /prism:doctor reports whether it’s on and the exact path.
What happens when the ingest service is unreachable?
Section titled “What happens when the ingest service is unreachable?”The submit hook fails open — your prompt still goes through. OTEL export retries internally. Prompt capture failures are logged to the debug log but not surfaced to you.
Scoring
Section titled “Scoring”How is my work scored?
Section titled “How is my work scored?”Prism looks at each sub-session and establishes four things: whether you reached the goal, whether the work had to be redone afterwards, how many tokens it took, and how well your prompts were written.
Those roll up into your Prompt Score — your average prompt grade, 0–100 with a letter grade — with the completion rate (the share of sub-sessions that crushed their goal) alongside it as a supporting number.
If the scoring can’t reach a confident verdict on a sub-session, it’s left unscored rather than guessed at.
What’s a good Prompt Score?
Section titled “What’s a good Prompt Score?”Your Prompt Score is your average prompt grade (0–100), shown with a letter grade on an A+ → F ladder. The per-prompt grade bands are:
Per-prompt grade: A+ 90–100, A 80–89, A− 72–79, B+ 62–71, B 50–61 (baseline), B− 42–49, C+ 34–41, C 24–33, D 12–23, F 0–11.
Your Completion rate — the share of sub-sessions you crushed — maps to a tier of its own. The Skill page has a ladder with the same five names at different cuts, so read the label beside it:
Completion tier (0–100):
- 90–100 — Elite
- 70–89 — Expert
- 50–69 — Proficient
- 30–49 — Practitioner
- 0–29 — Novice
A team’s completion rate is the team-aggregate crushed/total, never the average of per-user rates.
There’s no universal “good” number — track your own trend and focus on the prompt-rubric check you fail most often.
Do scores affect my Claude Code experience?
Section titled “Do scores affect my Claude Code experience?”No. Scores are informational. They don’t limit access, change model behavior, or affect billing.
How often are sessions scored?
Section titled “How often are sessions scored?”Scoring starts once a sub-session closes, and the results reach your dashboard within roughly 1–10 minutes. Nothing runs on a fixed daily schedule — if you’ve just finished something and don’t see it yet, give it a few minutes.
Are all seven rubric checks applied to every prompt?
Section titled “Are all seven rubric checks applied to every prompt?”No. Each prompt is first classified by what you were trying to do — ask a question, fix a bug, plan something, refactor, and so on — and only the checks that make sense for that kind of prompt are applied. Asking a plain question isn’t marked down for not requesting tests, for example. Checks that don’t apply are simply skipped, never counted against you.
Dashboard
Section titled “Dashboard”Why don’t I see any data?
Section titled “Why don’t I see any data?”Common causes:
- Plugin not configured — run
/prism:setup prism_YOUR_KEY. - Session too recent — wait 1–5 minutes for ingest, scoring, and analytics cache.
- API key issue — verify with
/prism:status. - OTEL export blocked by a firewall — telemetry ingestion needs outbound HTTPS to
ingest.optra-prism.com.
Can my team see my prompts?
Section titled “Can my team see my prompts?”Managers see team-level facts — what was worked on, usage, and cost. Your individual prompt text is never shown to them; your personal grades and coaching live in your own Daily Recap and Report.
How do I share insights with the team?
Section titled “How do I share insights with the team?”Unless you export it as a report, your insights are visible only to you. They reach your team in two ways:
- Export a report and share the file yourself. Personal reports export to Markdown, Word (.docx), or PDF.
- Team features share team-level facts with your team manager: in the Team Console a manager sees the team summary, performance, and cost. Access is set by organization role, and your individual prompt text, grades, and coaching are never exposed through it.