Agent skill · writing content · sickn33
interview-style-doc-building
Build structured strategy documents by asking one question at a time and patching the file.
Why this skill is useful
Provides a structured approach for building strategic documents through incremental user interaction, which the AI wouldn't reliably generate on its own.
What it needs
About 3k tokens when loaded. Last updated 2026-08-07. 44,571 stars on the source repository.
What this skill does
Interview-Style Doc Building The user's preferred mode for creating durable strategic docs. AI does NOT propose content — AI asks one question, the user answers, AI patches the file, AI asks the next question. The file IS the conversation's output, updated incrementally. When to Use Building a new SSOT file (life priorities, life vision, principles, frameworks, ranked lists). Filling out a structured doc the user explicitly wants to author themselves. Quarterly/annual reviews where the user's words go into the file. NOT for: day planning (use day-plan), task triage (organize-tasks), or anything where AI proposes content first. The Loop 1. Create the file with a skeleton (header, sections, "to be filled in" placeholders). Single writefile for the new file. After this, NEVER overwrite — only patch. 2. Ask ONE question. Concise. Specific. Single-faceted. Open-ended where possible. 3. Wait for the answer. Don't ask the next question yet. 4. Patch the file with the user's answer in the correct section. 5. Re-ask — next question, or follow-up if the answer was incomplete. 6. Repeat until the file is complete. Hard Rules One question at a time. Never dump multiple questions in a single message. The user has flagged this. Patch, don't overwrite. After the initial skeleton, use patch for every update. Never writefile to an existing doc. Update the file BEFORE asking the next question. Order: receive answer → patch file → ask next question. Not the reverse. Lists from the user are UNORDERED SETS. When the user lists items in response to "which X should we cover?" or "what are the Ys?", that is a SET, not a ranking. Never infer rank, priority, or sequence from the order they typed them. If you need ordering, ask explicitly: "Which of these is #1?" Ask dynamics, not names. When the user references a person, don't ask "who is X?" — ask about the role/dynamic. No snark, no attitude, no filler. Concise questions, concise acknowledgments. No speculative additions. …
How to use it
Reference it in AdaL, Claude Code, Cursor or any coding agent — nothing to install:
@skills sickn33/interview-style-doc-building--602d02