Usability Test Severity Codebook from Session Notes (No Invented Quotes)

Prompt

Act as a UX researcher building a severity codebook from usability session notes. You code usability findings with a severity scale the user pastes. You only quote notes that appear in Inputs. You do not invent quotes, task durations, or SUS scores. This is not an interview theme codebook and not a survey open-end codebook.
You work only from Inputs. Do not invent stats, citations, quotes, URLs, names, IDs, or records that are not in Inputs.

Inputs:
- Study name and prototype version: [Study]
- Tasks I actually ran: [Tasks]
- Severity scale I will use (paste definitions): [Scale]
- Pasted session notes (who, task, observation): [Notes]
- Participant labels as in the notes: [Who]
- Metrics I may use only if in Notes (time, success): [Metrics]
- Out of scope methods: [Out]
- Words I must not use: [Banned]
- Max findings: [Count or 8]
- Audience for the codebook: [Audience]

Generate:
1. Honesty ledger: study, tasks, scale definitions, participant labels, every quote-ready clause in Notes. Forbidden: quotes not in Notes, times not in Notes, SUS not in Notes.
2. Codebook table: code name, definition, severity from Scale, supporting quote (verbatim from Notes or NOT IN INPUTS), participant label from Who.
3. Cap at Count findings. Merge duplicates. Do not invent a P0 to make the study look dramatic.
4. Task map: which Task each finding came from. If Notes omit the task, write task unknown.
5. Metrics: only Metrics present in Notes. If no time was recorded, write time not recorded.
6. Out of scope: reject interview-life themes, survey codes, and methods in Out.
7. Gaps: five things the notes do not support (sample size claims, benchmark times).
8. Compliance pass: quote Banned words, invented quotes, fake SUS. Cut them.

Constraints:
- Usability-test codebook, not a qualitative interview theme codebook and not a survey codebook.
- Never invent a quote, participant, or task time.
- Severity labels only from Scale.
- Do not compute SUS or NPS unless Notes include the instrument.
- No emojis.

Instructions

Replace every [bracket] with your details before running. Works on ChatGPT, Claude, and Gemini.

Examples

Example Input

Study: Harbor Ledger iOS rec flow, prototype v4.2-rc1
Tasks: T1 find Rec; T2 complete a rec of a demo checking account; T3 export CSV
Scale: S1 cosmetic; S2 confusion recovered without help; S3 failed task or needed moderator; S4 data loss or cannot proceed. I pasted these four.
Notes: P3 T1 looked under Reports for Rec then found it under Accounts (no time). P3 T2 completed after moderator pointed at the red banner. P5 T3 CSV downloaded but the filename was export.csv and date unknown. P5 T2 success, no quote. No P4 notes pasted. No SUS.
Who: P3, P5 only
Metrics: no timers. Success/fail as in notes.
Out: interviews, surveys, NPS
Banned: 12 participants, 80 percent success, they said they loved it
Count: 6
Audience: iOS engineer plus CS

Example Output

1. Ledger. v4.2-rc1. Tasks T1 T2 T3. Scale S1-S4 as pasted. Who: P3, P5. Notes: P3 T1 Reports vs Accounts; P3 T2 needed moderator for red banner; P5 T3 export.csv filename. P5 T2 success no quote. No timers. No SUS. No P4. Forbidden: 12 participants, 80 percent, loved it, invented P4 quote.

2. Codebook. C1 Rec IA under Reports vs Accounts | S2 | P3 T1. C2 Rec completion needed moderator | S3 | P3 T2 observation, do not invent a quote. C3 CSV filename export.csv lacks date | S2 | P5 T3. P5 T2 success: no finding. Do not pad to 6 with invented S4 data loss.

3. Cap. Three findings. Count asked 6; notes support 3. Do not invent three more.

4. Task map. C1 T1. C2 T2. C3 T3.

5. Metrics. Time not recorded. Success: P5 T2 yes; P3 T2 no without help.

6. Out. No interview themes, no NPS, no survey codes.

7. Gaps. Sample size beyond P3 and P5; timers; SUS; P4; whether the red banner copy is known; whether filename is a spec.

8. Compliance. Cut banned words. Audience iOS engineer plus CS. Not an interview codebook.

Missing-data policy: if a field was blank, write NOT IN INPUTS rather than guessing. Lock any tool version named in Inputs; if unnamed, write unknown. No invented testimonials, star ratings, or press logos. If legal, clinical, insurance, HR, education-plan, or veterinary content appears, add a one-line not-advice and de-identify banner. Quote banned-word hits and cut them. End with a gaps list of five bullets the user still owes you. Character and byte caps in the job are hard; print counts when relevant. Refuse to backfill DOIs, exam dumps, PHI, PII, or compensation promises not in Inputs.