菜单
Synthesize a batch of customer interviews into evidence-for + evidence-against the hypothesis, with pattern surfacing and bias flagging.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Slice a goal into the smallest feature slices that can each be PROVEN — every slice ships with a named eval and an end-to-end test, and a slice you can't evaluate or test doesn't get selected. Use after the goal is set and before the spec, so only provable features make the build.
Guided, stateful provisioning of a product's production stack — deep-links, key-format validation, a secrets backend (fnox recommended, or user-touched .env.local), a resumable ledger, CLI automation after signup, and a blocking A1–A10 live gate. Hands off to /security-check → /ship. The walked version of SETUP.md.
Creates specs before coding. Use when starting a new project, feature, or significant change and no specification exists yet. Use when requirements are unclear, ambiguous, or only exist as a vague idea.
Launch the factory's control plane — scaffold and fill FACTORY-ORDERS (weekly mandate + hard budget), STANDING-ORDERS (autonomous-program authority), HEARTBEAT (weekly pulse), and write lifecycle gates for every registered product. A deliberate one-time ritual (re-run to audit); invoke as /factory-launch, not auto-triggered.
One-shot a new product — folders, Next.js starter, product.config.json, CLAUDE.md, scope.md, PRD skeleton, SETUP.md checklist. The Lovable-like ignition.
Breaks work into ordered tasks. Use when you have a spec or clear requirements and need to break work into implementable tasks. Use when a task feels too large to start, when you need to estimate scope, or when parallel work is possible.
| name | interview-synthesizer |
| description | Synthesize a batch of customer interviews into evidence-for + evidence-against the hypothesis, with pattern surfacing and bias flagging. |
| model_tier | sonnet |
User has notes from 5+ interviews and asks: "synthesize these", "what did I learn from these calls?", "are we seeing PMF signals?"
Saved to products/<name>/interviews/synthesis-<date>.md:
## Synthesis — interviews <date range> (N conversations)
### Evidence FOR the hypothesis
- <quote from interview> — (P3 — Sarah, Senior PM)
- <quote> — (P5 — ...)
- ...
### Evidence AGAINST the hypothesis
- <quote> — (P2)
- <quote> — (P7)
- ...
### Surprises (what we didn't expect)
- <observation> — surfaced by P4, P6
- ...
### Patterns
- N of N interviewees mentioned <specific behavior / phrase>
- <segmentation pattern>: <subgroup> behaves differently from <other subgroup>
### The strongest single quote
> "<verbatim>" — Person, Role
### Bias check
Compare lists:
- Evidence FOR: <count>
- Evidence AGAINST: <count>
- Ratio: <FOR/AGAINST>
- If FOR >> AGAINST, ask: is this real, or are we hearing what we want? Specifically: did our questions lead the witness? Did we ignore disconfirming signals?
### Recommendation
- CONTINUE: validation pattern is strong — refine and run 5 more
- PIVOT: the right problem nearby is <X>; rerun discovery
- KILL: evidence against is overwhelming — log learnings, move on
- INCONCLUSIVE: too noisy — change target profile or sharpen hypothesis
products/<name>/interviews/raw/).Per factory/playbooks/mvp-stage/agent-handoff-contracts.md:
products/<name>/interviews/raw/ or user-passed paths), each attributable to a distinct person matching the discovery plan's target profile.products/<name>/interviews/synthesis-<date>.md in the exact format above.factory/playbooks/idea-stage/mom-test.mdfactory/playbooks/ai-native-2026/founders-playbook-distilled.md (re: bias check)