borghei/Claude-Skills / research-ops/product-research
product-research
VerifiedContinuous product discovery research operations — picking the method that fits the question, recruiting and screening participants, building interview guides, and scoring evidence into insights. Use when planning or running discovery.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 13.6 KB |
| assets/interview-guide-template.md | markdown | 4.6 KB |
| assets/sample_evidence.json | config | 2.8 KB |
| assets/sample_research_question.json | config | 394 B |
| assets/sample_screener.json | config | 2.4 KB |
| references/interview-craft.md | markdown | 17.1 KB |
| references/method-selection-guide.md | markdown | 16.7 KB |
| scripts/insight_confidence_scorer.py | script | 9.9 KB |
| scripts/method_recommender.py | script | 11.7 KB |
| scripts/screener_validator.py | script | 12.3 KB |
assets/interview-guide-template.md
View on GitHub# Interview Guide — [Study Name]
**Decision this informs:** [the specific decision and who makes it]
**Research question:** [interrogative form]
**Segment:** [which segment this session covers] | **Session:** [n of N]
**Moderator:** [name] | **Notetaker:** [name] | **Date:** [YYYY-MM-DD]
**What result would change our course:** [Write this before the first session.
If you cannot, the decision is not actually open and this study is decoration.]
**Target length:** 45 minutes
---
## 0. Opening — verbatim (2 min)
> Thanks for making time. I work on [product] and I'm trying to understand how
> [work area] actually happens day to day — not to demo anything or get feedback
> on a design. There are no wrong answers; I'm interested in what you actually
> do, including the parts that are messy.
>
> I'd like to record this so I don't have to take notes while we talk. The
> recording goes to [who], is used for [purpose], and is deleted after [period].
> It won't be shared outside that. Is that OK?
>
> You can skip any question or stop at any point. Any questions before we start?
- [ ] Consent obtained and recorded
---
## 1. Warm-up (3-5 min)
Goal: get them talking in specifics; capture the context for everything after.
- Walk me through what a normal [Tuesday] looks like for you.
- How long have you been doing this? What were you doing before?
- Who else is on your team, and how is the work divided?
---
## 2. Context (5-8 min)
Goal: establish the problem space in **their** vocabulary, before introducing any
of yours.
- Where does [the work] come from, and who hands it to you?
- What tools are involved from the moment it arrives to the moment it's done?
- What's changed about that in the last year?
**Their vocabulary — capture verbatim and use it for the rest of the session:**
| They say | We say |
|----------|--------|
| [ ] | [ ] |
| [ ] | [ ] |
---
## 3. Specific-instance core (20-30 min)
Goal: concrete past occurrences. This is where the value is. Never accept
"usually" — always convert to "the last time."
**Must ask:**
- Tell me about the last time [the situation under study] happened.
**Walk the timeline:**
- What triggered it? How did you notice?
- What did you do first?
- What did you try after that?
- When that didn't work, what then?
- Who else got involved?
- How did it end?
- How long did the whole thing take?
- What did it cost — your time, the customer, anything else?
**Then:** Was that a typical instance, or unusual? Tell me about one that went
differently.
**Probes available:** silence (3-5 seconds) · "Tell me more about that." ·
"What happened next?" · "Why does that matter to you?" ·
"You mentioned [X] — say more?" · "How did you learn to do it that way?" ·
"What would you have done if that hadn't worked?"
---
## 4. Workaround excavation (5-10 min)
Goal: find where someone cared enough to build an unsupported solution.
- Is there anything you do outside the system to make this work?
- Do you keep anything in a spreadsheet, notes app, or on paper?
- Has anyone on the team built something for this themselves?
- What do you check before you trust what the system tells you?
- Is there anything you've stopped using because it wasn't worth it?
---
## 5. Prioritisation (5 min)
Goal: force the trade-off. Unranked lists come back with everything important.
- Of everything we've discussed, which one — if it disappeared tomorrow — would
change your week the most?
- If you could fix exactly one and the others stayed as they are, which?
- What would have to be true for you to change how you do this?
---
## 6. Close (2-3 min)
- What should I have asked you that I didn't?
- Who else should I be talking to about this?
- Would you be open to a follow-up if something comes up?
> Thank you — this was genuinely useful. [Incentive logistics.]
---
## Notetaker sheet
**Topics covered:** ☐ warm-up ☐ context ☐ specific instance ☐ workarounds ☐ prioritisation ☐ close
**Observations** — one atomic observation per row, no interpretation:
| # | Observation | Kind (observed / artefact / reported) | Timestamp |
|---|-------------|---------------------------------------|-----------|
| 1 | | | |
| 2 | | | |
| 3 | | | |
**New themes this session** (for saturation tracking): [list — if zero for two
consecutive sessions in this segment, stop recruiting]
**Contradicted an earlier session?** [what, and which session]
**Guide changes for next session:** [what to add, drop, or reorder]
---
## Debrief — complete within 15 minutes of the session
- **Top three things heard:**
1.
2.
3.
- **Biggest surprise:**
- **Moderator airtime — under 20%?** ☐ yes ☐ no
- **Anything we led the participant toward?**
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| add-product-e2e-evalpaperclipai/paperclipAdd or extend a Paperclip full-stack runner E2E workflow, fixture, matcher, or report evidence path for local or Daytona execution. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |
| add-runner-evalpaperclipai/paperclipAdd or extend a Paperclip Runner protocol evaluation definition, roster, assertion, or report fixture with provenance and narrow validation. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |
| diagnose-why-work-stoppedpaperclipai/paperclipCompatibility entry point for failed, stalled, looping, or over-recovered Paperclip work. Use fix-failed-paperclip-run for the full diagnosis, regression, repair, real verification, and green-PR workflow; honor an explicitly requested diagnosis-only scope. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |
| fix-failed-paperclip-runpaperclipai/paperclipDiagnose and fix failed, stuck, slow, or incorrectly blocked Paperclip tasks and runs. Use with a task/run URL or failure report to trace the root cause, prove a regression red then green, check second-order effects, verify the real workflow, and deliver a PR with passing checks. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |