core-interview — Claude Code skill: what it does, how to install
core-interview — a skill for Claude Code that runs an interview loop through decision trees until every branch is settled.
MIT 1,000 installs ★ 10 updated 2026-10-16
npx skills add pwguler/skills --skill core-interview -a claude-codeWhat core-interview does
The skill activates when triggered by the user or another skill to guide requirements gathering step by step. It evaluates the codebase before asking questions, adopts recommendations when choices carry low risk, and groups low-risk clarifications together. Reference implementations are requested when the target behavior is hard to describe in words.
A final summary lists all adopted assumptions, and rejecting any assumption reopens only its specific branch for revision.
What it helps with
- settle requirements through a decision tree
- gather missing details before building
- look up facts in the codebase instead of asking
- compile a summary of adopted assumptions
Example prompts for the agent
- Run an interview for the new authentication module
- Ask questions about the database schema step by step
- Start the core-interview loop for this feature
How to install core-interview
1Skills CLI (Vercel)
npx skills add pwguler/skills --skill core-interview -a claude-code2Manually: put the files in place
~/.claude/skills/core-interview/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/core-interview/
1Skills CLI (Vercel)
npx skills add pwguler/skills --skill core-interview -a codex2Manually: put the files in place
.agents/skills/core-interview/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add pwguler/skills --skill core-interview -a cursor2Manually: put the files in place
.cursor/skills/core-interview/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install pwguler/skills skills/core-interview2Manually: put the files in place
.github/skills/core-interview/SKILL.md1Skills CLI (Vercel)
npx skills add pwguler/skills --skill core-interview -a gemini-cli2Platform CLI
gemini skills install https://github.com/pwguler/skills --path skills/core-interviewManually: put the files in place
.devin/skills/core-interview/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add pwguler/skills --skill core-interview -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about core-interview
How to install core-interview in Claude Code?
Run npx skills add pwguler/skills --skill core-interview -a claude-code and check ~/.claude/skills/core-interview/SKILL.md.
What does core-interview do?
It interviews the user and checks the codebase to resolve every branch of a task before code generation starts.
When does this skill run?
It runs exclusively when started by the user or another skill, never on its own.
Source text
SKILL.md · MIT
--- name: core-interview description: Internal support skill that runs the interview loop for other skills, until every decision is settled. Use it only when another skill or the user starts it, never on your own. --- Question the user until the two of you agree on every part of the topic. Treat the topic as a decision tree. Settle each decision before the decisions that depend on it. Leave no branch open. Not every branch needs a question to the user. The code can settle it, or an assumption that the rules below allow. Rules: - Ask only questions whose answer would change what gets built. When your recommendation is strong and a wrong choice costs little, do not ask. Adopt the recommendation, record it as an assumption, and continue. - Put one real decision at a time to the user. Use the question tool of your agent when it has one. Wait for the answer before you ask the next question. The options are your candidate answers, with the recommended answer first. Put low-risk clarifications together in one question. - The user can always type an answer that the options do not have. So the tool never limits the choices of a decision. - If the codebase can answer a question, look in the codebase. Do not ask the user. You find the facts, not the user. Look up a fact in this session before you ask the question that depends on it. - When the user cannot describe the target behavior in words, ask for a reference implementation. That is source code whose behavior you reproduce, also in another language or library. Source code is better than a description or a screenshot. - The final summary lists every assumption that you adopted. The user can then reject any of them in one reply. A rejected assumption opens only its own branch again.
© pwguler · MIT · pwguler/skills