answer-question

Read report.md and results.csv to answer questions and save answer.md with meta.json.

Updated Feb 5, 2026
One-click install
npx skills add https://github.com/expectedparrot/skills --skill answer-question
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: answer-question
Source: https://github.com/expectedparrot/skills/tree/main/.claude/skills/answer-question
Command: npx skills add https://github.com/expectedparrot/skills --skill answer-question

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Answer questions about a generated analysis report - reads report artifacts, performs additional analysis if needed, and saves the answer with metadata

Core Features & Use Cases

  • Reads analysis reports (report.md) and data (results.csv) to answer user questions.
  • Optionally runs reproducible Python analyses (analysis.py) when needed and saves artifacts under a per-question output directory.
  • Generates a markdown answer (answer.md) and a metadata file (meta.json) for traceability.

Quick Start

Provide a question and optional analysis directory to generate a structured answer from the most recent analysis run.

Frequently Asked Questions about answer-question

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I answer questions about a markdown analysis report using CSV data?

To answer questions about a markdown analysis report, the system reads report.md and results.csv data to generate a structured answer. If the existing report is insufficient, it runs optional deeper analysis and saves artifacts in a per-question output directory.

What is the best way to extract specific data insights from a generated analysis report?

Extract data insights from an analysis report by querying the system to read the existing report and results.csv. It performs additional reproducible Python analyses when needed and returns a structured markdown answer with metadata for traceability.

Can I run reproducible Python analysis if my current report cannot answer my question?

Yes, if the existing report cannot fully answer your question, the system runs optional reproducible Python analyses. It executes the analysis and saves any resulting artifacts in a per-question output directory for reference.

Does this question-answering approach save metadata for traceability with the answers?

Yes, the question-answering process saves metadata for traceability. It always writes a structured markdown answer file alongside a metadata file to ensure every generated response is traceable to its source analysis and data.

What do I need to generate a structured answer from an existing analysis run?

To generate a structured answer from an analysis run, provide a specific question and an optional analysis directory. The system reads the latest report.md and results.csv artifacts to construct the answer and metadata.