Install
$ agentstack add skill-kellykampen-agent-skills-interview ✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
Verified badge
Passed review? Show it. Paste this badge into your README, it links to the public security report.
Reliability & compatibility
Declared compatibility
Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.
We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.
How agent discovery & health will work →About
Interview
Interview the user thoroughly, then persist the result. The goal is to close the gap between what they have in their head and what's actually written down — surfacing decisions, tradeoffs, and edge cases they haven't thought about yet.
How to interview
- Use the AskUserQuestion tool for every question — never a wall of prose questions. Batch up to 4 at a time, with smart multiple-choice options plus room for free-text.
- Ask about anything that matters: technical implementation, UI/UX, data model, edge cases, failure modes, scope, concerns, tradeoffs, non-goals.
- Make the questions non-obvious. Don't ask what you could infer from the code or the request. Ask the things that would actually change the design.
- If you're in a git repo, read the relevant codebase first and ask questions grounded in it (existing patterns, conventions, where this fits).
- Keep going — interview continually, round after round, until the picture is genuinely complete. Don't stop early.
Modes — pick based on what the user asked for
1. Plan file (default)
/interview — read the given plan/spec file, interview to fill its gaps, then write the finished spec back to that same file.
2. Feature → plans/
When interviewing about a feature or refactor (a description, a comment, a ticket), write the finished spec to ./plans/.md (create the plans/ dir if needed). Ground the questions in the current codebase.
3. Feature → Linear
When the user wants issues created, after the interview use the linear-* skills to create:
- A Linear project capturing the what / why / how / key features.
- Issues under the repo's team (usually named similarly to the repo), each ≤ 3 points of effort — break anything larger down.
- Links between issues for blockers / dependencies.
Finish
State clearly where you wrote the spec (file path or Linear project/issue links) and give a short summary of the key decisions the interview locked in.
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: kellykampen
- Source: kellykampen/agent-skills
- License: MIT
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet, be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.