# Interview

> Extract process knowledge from an SME through structured MECE interview. Use when user says "interview me about a workflow", "help me map this process", "extract process knowledge", or wants to build a decomposition through guided conversation.

- **Type:** Skill
- **Install:** `agentstack add skill-fblissjr-fb-claude-skills-interview`
- **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior
- **Seller:** [fblissjr](https://agentstack.voostack.com/s/fblissjr)
- **Installs:** 0
- **Category:** [Agent Skills](https://agentstack.voostack.com/c/agent-skills)
- **Latest version:** 0.1.0
- **License:** Apache-2.0
- **Upstream author:** [fblissjr](https://github.com/fblissjr)
- **Source:** https://github.com/fblissjr/fb-claude-skills/tree/main/apps/mece-decomposer/skills/interview

## Install

```sh
agentstack add skill-fblissjr-fb-claude-skills-interview
```

Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent.

## About

# /interview

Structured conversational extraction of process knowledge from a subject matter expert. Produces a MECE decomposition through guided dialogue.

## Usage

```
/interview 
```

Examples:
- `/interview our invoice approval workflow`
- `/interview the deployment process for our SaaS platform`
- `/interview` then describe what process you want to map

## How It Works

```
STANDALONE (always works)
- Structured 5-phase interview protocol
- Adapts to your communication style
- Produces MECE tree + Agent SDK JSON

SUPERCHARGED (when MCP server connected)
+ Interactive tree visualization as you build it
+ Click nodes to refine, validate scores live
```

## Interview Protocol

### Phase 1: Context and Scope (3-5 questions)

Establish what we are decomposing:
- What is the process and who is involved?
- What triggers it and when is it done?
- What is explicitly excluded?

### Phase 2: Happy Path Walkthrough (open-ended)

You narrate the end-to-end process as it works when everything goes right. I probe each step for:
- Who does it (actor)
- How long it takes (duration)
- What goes in and what comes out (inputs/outputs)
- Is it conditional or always happens (conditionality)
- Can things happen in parallel (parallelism)

### Phase 3: Exception Discovery (per step)

Go back through each step and surface:
- What fails and how
- Edge cases and skip conditions
- Branching logic

### Phase 4: Boundary Conditions

Discover handoffs, parallel activities, upstream dependencies, downstream consumers, time constraints, and approval gates.

### Phase 5: Validation

Present the decomposition tree and iterate until you confirm it covers the process completely without overlap.

## Adaptive Behavior

I adjust interview style based on how you think:

- **Structured thinker**: I match your top-down approach, validate MECE against your mental model
- **Narrative thinker**: I let you tell stories, organize afterward, use your stories as validation scenarios
- **Time-constrained**: I show a draft decomposition and let you correct it (Phase 5 first)

## Output

Same dual output as `/decompose`:
1. Human-readable tree (markdown)
2. Agent SDK JSON
3. Interactive visualization (when MCP server connected)

## Tips

1. **Don't worry about structure** -- just describe how things work. I'll organize it.
2. **Include the messy parts** -- exceptions and edge cases are the most valuable.
3. **Say "it depends"** -- that's a branching point, and I'll probe for conditions.
4. **Correct me freely** -- the tree is wrong until you say it's right.

## Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

- **Author:** [fblissjr](https://github.com/fblissjr)
- **Source:** [fblissjr/fb-claude-skills](https://github.com/fblissjr/fb-claude-skills)
- **License:** Apache-2.0

Install and usage instructions live in the source repository linked above.

## Pricing

- **Free** — Free

## Security capabilities

Automated source analysis of v0.1.0 — what this tool can access:

- **Network access:** no
- **Filesystem access:** no
- **Shell / process execution:** no
- **Environment & secrets:** no
- **Dynamic code execution:** no

*"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.*


## Versions

- **0.1.0** — security scan: passed — Imported from the upstream source.

## Links

- Listing page: https://agentstack.voostack.com/l/skill-fblissjr-fb-claude-skills-interview
- Seller: https://agentstack.voostack.com/s/fblissjr
- Browse the marketplace: https://agentstack.voostack.com/browse

---
Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.