| name | create-questions-and-answers-en |
| version | 1.0.0 |
| description | Runs a structured Q&A loop with a user: prompts them for questions one at a time, answers each question, then compiles all questions and answers into a downloadable Markdown file at the end. Use this skill whenever a user wants to test a configuration, probe knowledge, document a session, or explore a topic interactively. Activate for phrases like: "let us do a Q&A", "ask me questions", "I have questions for you to answer", "test my setup with questions", "compile a Q&A document", or "end of questions".
|
| license | CC BY-NC-SA 4.0 |
| language | en |
| author | roebi (Robert Halter) |
| attribution | Part of the roebi agent-skills library. https://github.com/roebi/agent-skills
|
| spec | https://agentskills.io/specification |
| compatibility | Requires file output capability (bash or file_create tool). |
| metadata | {"author":"roebi","spec":"https://agentskills.io/specification"} |
Create Questions and Answers
Guides an agent through a structured, iterative Q&A session with a user and
produces a downloadable Markdown file containing all questions and answers.
When to use this skill
Activate when the user:
- Wants to test a system prompt, configuration, or knowledge base
- Wants to document a topic through guided questions
- Asks for a Q&A loop, question-answer session, or interview-style interaction
- Mentions "end of questions" as a termination signal
- Wants a compiled, downloadable record of the conversation
Step-by-step instructions
1. Confirm understanding
At the start, briefly confirm the loop to the user:
- You will prompt them to supply questions one at a time
- You will answer each question
- The loop repeats until the user says "end of questions"
- You will then compile everything into a downloadable
.md file
2. Run the Q&A loop
Repeat this cycle until termination:
1. Prompt the user: "Ready — go ahead with your next question."
2. User writes the question.
3. Agent answers the question thoroughly and accurately.
4. Return to step 1.
Rules during the loop:
- Answer each question before asking for the next one — never batch questions
- Use web search or tools if the question requires current or external information
- Cite sources where relevant
- Keep answers focused; avoid padding
3. Detect termination
The loop ends when the user writes any of:
end of questions
no more questions
that's all
done
4. Compile the Markdown file
Once the loop ends, create a .md file containing:
# Q&A Session: <topic or session title>
> <optional: brief description of the session context>
> Date: <YYYY-MM-DD>
---
## Q1: <Short question title>
**Question:**
<Full question text>
**Answer:**
<Full answer text>
---
## Q2: <Short question title>
...
File naming convention:
<topic-slug>-qa-session.md
# Examples:
# systemprompt-qa-session.md
# agent-skills-qa-session.md
# project-review-qa-session.md
5. Deliver the file
- Save the file to the output directory
- Present it to the user using the
present_files tool
- Provide a one-sentence summary of how many questions were covered
Output format rules
- Use
## Q<N>: <title> headings for each entry
- Bold
**Question:** and **Answer:** labels
- Separate entries with
---
- Preserve the full original question text verbatim
- Preserve the full answer including any tables, lists, or code blocks
- Include a session header with topic and date
Examples
Good session opener
"Understood! Here's the loop: you give me a question, I answer it, we repeat
until you say 'end of questions', then I compile everything into a .md file.
Ready — go ahead with your first question!"
Good termination handling
User: end of questions
Agent: (immediately stops the loop, compiles the file, presents it)
"All N questions compiled into <topic>-qa-session.md."
Anti-patterns to avoid
- ❌ Asking multiple questions at once
- ❌ Answering before the user has submitted their question
- ❌ Skipping the compiled file output
- ❌ Truncating long answers in the output file
- ❌ Forgetting the session date in the header
Reference files in this skill
references/output-format.md — extended Markdown output formatting guide