- Home
- Skills
- Agents & Orchestration
- Problem Intake Interviewer (beta)
More screenshots
Works with the AI Agents you already use
Problem Intake Interviewer (beta)
Transform vague automation ideas into rigorous, builder-ready problem charters through structured interviews.
Free
Problem Intake Interviewer (beta)
Example session with this skill installed
I want to build an agent that summarizes my daily meetings from Zoom transcripts and posts them to Slack. Help me define this.
- Read your context and instructions
- Compiled the problem intake interviewer
I will help you define this project. We will go through a few rounds of at most three questions each, followed by a short playback to confirm the logic, and finally, a formal problem charter. "Don't know yet" is always a valid answer.
First, I need to check two things:
1.
Data Sharing: Will these transcripts contain confidential, personal, or customer data? If so, we should use public or anonymized samples for this scoping exercise.
2.
Size: This sounds like a single-purpose agent for summarizing and posting. Is this intended for just your team, or is this a platform for an entire organization?
Round 1
Goal (Stage 1): You mentioned "summarizing daily meetings." Does this refer only to your own meetings, or all meetings occurring in a specific Zoom account? [A01]
Output (Stage 1): You mentioned "posts them to Slack." Should the summary include specific sections like action items, attendees, or sentiment, or just a general paragraph? [A01]
Human-in-the-loop (Stage 2): Since this writes to Slack, where should the human approval gate happen? Should it wait for your approval in a private DM before posting to a public channel?
Next steps
- Answer the qualification questions regarding data and size.
- Provide details on the specific output format you expect in Slack.
- Confirm if you have a sample transcript file we can review.
Connects securely to your tools. The creator never sees your data.
About this skill
Vague project goals and one-line prompts lead to failed agents and wasted compute. This skill interviews problem owners to transform fuzzy ideas into rigorous problem charters that a developer or another agent can build against without guesswork.
What it does
- Qualifies scope by identifying sensitive data risks and ensuring the project size is manageable for a single system or script.
- Scans requirement gaps by auditing missing, vague, or contradictory information against a structured question bank.
- Probes for clarity using targeted techniques to resolve ambiguity without leading the owner or suggesting answers.
- Performs SCQA playback to confirm the situation, complication, question, and answer in plain language before committing to a spec.
- Generates a full charter complete with goal, scope, quality rules, severities, and human-in-the-loop approval gates.
How it works
- Initialize by reviewing any existing briefs or sample files and outlining the interview process.
- Qualify the data privacy requirements and project scale to ensure the scope is appropriate.
- Interview the owner in short rounds, citing every answer and file to maintain a clear trail of intent.
- Confirm the logic via a Situation-Complication-Question-Answer playback.
- Document the final charter using the confirmed playback, including specific success and failure examples.
Frameworks & tools
The skill is framework-agnostic and relies on structured Markdown templates. It utilizes a question-bank.md for requirement gathering and probes.md for systematic clarification.
Why this beats prompting it yourself
Generic prompts often lead the user or fill in missing details with LLM hallucinations, creating a spec the user didn't actually intend. This skill enforces a "never fill the gap" rule, ensuring every requirement is owner-verified and every unknown is logged as an open question.
Use cases
- Scoping an agent to handle specific data extraction tasks with strict quality rules.
- Writing a PRD for a new AI feature where inputs and outputs are not yet defined.
- Automating a workflow that currently requires manual human coordination and approval.
- Defining acceptance tests for a script to ensure it handles edge cases correctly.
Known limitations
The skill is limited to five interview rounds by default to prevent scope creep. It is designed for single systems or small agent teams, not for enterprise-wide platform architectures.
Limited to 5 interview rounds by default. Not suitable for multi-team programs or enterprise platforms. Requires human confirmation of playback before generating the final charter.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Free forever
No account required to browse
Trust & safety
Security scan
Security scan passed 11 days ago
- Free with an Agensi account