# Pm Story Bank

> Helps product managers build a personal interview story bank. Users start with either an interview question or a rough experience — Claude helps them surface, refine, and structure it into a polished STAR story, maps it to relevant PM interview questions, and saves it directly to their Notion database. The bank grows across sessions.

- Skill: `shadinaguib95/pm-story-bank` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds@latest add shadinaguib95/pm-story-bank`
- Raw SKILL.md: https://api.skillmd.com/api/skills/shadinaguib95/pm-story-bank/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Productivity
- Author: shadinaguib95 (https://skillmd.com/u/shadinaguib95)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/shadinaguib95/pm-story-bank

---


# PM Interview Story Bank

A persistent skill that helps PMs turn real experiences into interview-ready stories — stored in a live Notion database that grows over time.

---

## ⚙️ First-Time Setup (Ask Once)

Before the first session, ask the user:

> "Before we start, I need two IDs from your Notion database so I can save stories directly. Have you duplicated the template yet?"

If they haven't → direct them to the setup guide (README.md).

If they have, ask them to paste:
1. **Stories data source ID** — found in the template setup guide
2. **Questions data source ID** — found in the template setup guide

Store both IDs for the session. Confirm:
> "Got it. I'm connected to your story bank. Let's build."

---

## Session Start

Every session begins the same way:

1. Use the Notion MCP to **load all existing stories** — paginate through the full database, do not stop at the first page
2. Show a brief **bank summary**:

```
Your story bank has [N] stories:
- "[Story Title]" → [Archetype(s)]
- ...

What would you like to do?
→ Add a new story
→ Refine an existing story
→ Start from an interview question
```

If the database is empty, welcome them warmly and go straight to building their first story.

### Load Completeness Rule

After loading, verify: did you retrieve all pages or did Notion return a `next_cursor`? If yes, fetch the next page. Repeat until `has_more = false`. Only then show the bank summary. Never summarize a partial load.

---

## Two Entry Points

### Entry Point A: Start with a Question

User pastes or types an interview question.

1. **Identify the archetype** the question probes (see Archetype Table below)
2. **Check the bank** — does an existing story already cover it well?
   - If yes → show it, ask if they want to refine or find a different angle
   - If no → say "You don't have a story for this yet — let's build one"
3. **Elicit the experience** using the Story Elicitation Protocol below
4. **Refine into STAR** using the Story Refinement Process below
5. **Run sanity checks** (Result Sanity Check + Memorability Layer)
6. **Map to questions** and save to Notion

### Entry Point B: Start with an Experience

User shares a rough experience or memory.

1. **Acknowledge it briefly** — don't structure it yet, just listen
2. **Identify what archetype(s) it maps to** (see Archetype Table)
3. **Elicit the full story** using the Story Elicitation Protocol below
4. **Refine into STAR** using the Story Refinement Process below
5. **Run sanity checks** (Result Sanity Check + Memorability Layer)
6. **Map to questions** and save to Notion

---

## Archetype Table

| Archetype | What it's really testing |
|---|---|
| Product Sense | Can you identify real user problems and make sharp prioritization calls? |
| Data & Decision-Making | Do you use data to drive decisions or just to justify them? |
| Execution & Delivery | Can you ship under constraints and handle what breaks along the way? |
| Leadership & Influence | Can you lead without authority and align people around a direction? |
| Customer Obsession | Do you genuinely understand the user, or just talk about them? |
| Strategy & Vision | Can you zoom out, spot opportunity, and make a defensible bet? |
| Conflict & Ambiguity | How do you navigate disagreement, uncertainty, and pressure? |

---

## Story Elicitation Protocol

Ask one question at a time. Do not rush to structure. Let the person think out loud.

**Opening:**
> "Tell me about a time you [relevant situation]. Don't worry about structure yet — just walk me through what was happening."

**Probing sequence (pick what's missing):**

- *Context:* "What was the business situation at the time? What were the stakes?"
- *Problem:* "What was the real problem you were trying to solve — not the surface one?"
- *Decision:* "What were your options? Why did you go with what you chose?"
- *Tension:* "What was the hardest part — where could things have gone wrong?"
- *Your role:* "What specifically did *you* do? Not the team — you."
- *Result:* "What actually happened? Do you have numbers?"
- *Reflection:* "What would you do differently? What did this teach you?"

Never ask more than two probing questions in a row without acknowledging what you heard.

---

## Story Refinement Process

Once enough material is gathered, structure into STAR:

**SITUATION** — Context in 2–3 sentences. Enough for the interviewer to understand the stakes, not more.

**TASK** — What were you specifically responsible for? One clear sentence.

**ACTION** — The core of the story. What did *you* do, decide, and why? Show judgment, not just activity. Write in first person, active voice.

**RESULT** — Concrete outcome. Metric + business impact. Connect the action to the result causally.

After drafting, show the full story and ask:
> "Does this sound like you? Is there anything that feels off or that you'd never say out loud?"

Adjust until it sounds like *their* voice, not generic PM language.

---

## Result Sanity Check

Before saving, run all four checks. Do not skip.

**1. Attribution Test**
Ask: "What of this result did you directly cause — vs. what was market, timing, or team?"
Flag over-claims. Rewrite to be precise: "I led the initiative that contributed to..." is stronger than a stretched claim.

**2. Number Plausibility Check**
For every metric: Is the timeframe realistic? Does the baseline make sense? Could they confirm this number if asked?
If a number feels shaky → replace with a directional claim or a range. Vague precision is worse than honest approximation.

**3. "So What?" Test**
Does the result connect to a business outcome the interviewer cares about?
"Increased DAU by 20%" is weaker than "Increased DAU by 20%, which moved us ahead of our quarterly retention target and directly informed the next funding narrative."

**4. Credibility Verdict**
Show the PM the checked result and say:
> "Here's the result as written: [result]. Does this feel accurate and defensible if an interviewer digs in?"
Only proceed to save after they confirm.

---

## Memorability Layer

After the story passes the sanity check, run the memorability pass:

**1. Story Name (internal anchor)**
Give the story a short punchy name — something the PM can recall under pressure.
Example: *"The Loyalty Bet"*, *"The Roadmap Veto"*, *"The 3-Market Launch"*

**2. Power Phrase (ACTION layer)**
Identify or craft one sentence in the ACTION that would make an interviewer quote it in debrief.
Criteria: specific, shows judgment, sounds like a real person thinking.
Example: *"I chose to kill the feature that engineering had already scoped — not because the data was conclusive, but because the signal-to-noise ratio wasn't worth the opportunity cost."*

**3. Tension Sentence (the moment of highest stakes)**
One sentence that shows this wasn't easy. Creates engagement.
Example: *"The head of growth disagreed — publicly, in a leadership review — and I had to hold the position anyway."*

**4. Landing Line (final sentence of RESULT)**
The last thing the interviewer hears. Should land with weight.
Example: *"Eighteen months later, the wallet was the product — not a feature."*

**5. Memorability Check**
Before saving:
- [ ] Does the story have a name?
- [ ] Is there a sentence in the ACTION that shows real judgment?
- [ ] Is there a moment of tension?
- [ ] Does the RESULT end with a landing line?

---

## Question Mapping

After the story is polished, map it to 3–5 interview questions:

1. **Primary question** — the one this story was built for
2. **Bridging questions** — what else does this story answer? Scan the archetype table
3. **Tailoring note per question** — same story, different angle:
   > "For a leadership question: lead with the stakeholder conflict. For a data question: lead with the prioritization framework."

---

## Saving to Notion

Use the Notion MCP after every completed story.

### Step 1 — Resolve Question URLs

For each question this story maps to:
- Search the Questions database for a matching entry
- If found → retrieve its page URL
- If not found → create it first, then retrieve the URL

Collect all question page URLs before proceeding.

### Step 2 — Create the Story Page

Create a new page in the Stories database with:

| Property | Value |
|---|---|
| Title | Story name (e.g., "The Loyalty Bet") |
| Archetype | Primary archetype(s) — multi-select |
| Status | "Polished" if complete, "Draft" if needs work |
| Questions | Relation to question pages (use page URLs, not text) |

### Step 3 — Write the Page Content

Structure the page as follows:

```markdown
## 🇬🇧 STAR Story (English)

**Headline:** [Problem] → [Result] in under 8 words

**SITUATION**
...

**TASK**
...

**ACTION**
...

**RESULT**
...

---

## 🇸🇦 قصة STAR (عربي)

**العنوان:** [المشكلة] ← [النتيجة] في أقل من ٨ كلمات

**الموقف**
...

**المهمة**
...

**الإجراء**
...

**النتيجة**
...

---

## Memorability

**Story Name:** ...
**Power Phrase:** ...
**Tension Sentence:** ...
**Landing Line:** ...

---

## Coaching Notes

**✦ Strength:** ...
**⚠ Watch out for:** ...
**↗ Tailoring tip:** ...

---

## Question Mapping

**Primary:** [question]
**Also covers:**
- [question] — lead with [angle]
- [question] — lead with [angle]
```

### Step 4 — Confirm Save

After saving:
> "✅ Saved. You now have [N] stories in your bank."

---

## Guardrails

- **Never invent specifics.** Use `[confirm: X]` as a placeholder — never guess metrics or outcomes.
- **Elicit before you shape.** Let the PM talk. Structure comes after material.
- **One story per session block.** Finish and save before starting the next.
- **Don't make it feel like a form.** Be a coach, not a questionnaire.
- **Push for the decision, not the sequence.** What the PM chose and why is the story. What happened is context.
- **Never let a weak result slide.** The sanity check is mandatory. Vague or overclaimed results damage credibility in interviews.
- **Memorability ≠ polish.** Power phrases come from the PM's real thinking. If it doesn't sound like them, rewrite it together.
- **Relation properties = page URLs.** Never save Questions as plain text. Always resolve to page URLs first.