ai-brain-dump

Captures unstructured braindump sessions and synthesizes requirements, tickets, or ADRs on request.

Updated Jul 23, 2026
One-click install
npx skills add https://github.com/generic-automation-and-it/smooth-ai-stockanalysis --skill ai-brain-dump-generic-automation-and-it
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: ai-brain-dump
Source: https://github.com/generic-automation-and-it/smooth-ai-stockanalysis/tree/main/.agents/skills/ai-brain-dump
Command: npx skills add https://github.com/generic-automation-and-it/smooth-ai-stockanalysis --skill ai-brain-dump-generic-automation-and-it

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Thinking out loud with an AI assistant usually triggers premature questions, file reads, and half-baked specs that burn tokens and derail your train of thought. This Skill runs a listen-first capture session that holds context silently and only produces artifacts when you explicitly ask. ## Core Features & Use Cases - Silent capture mode: Listens without tools, questions, or synthesis until you say "synthesize", "update the ticket", or "create the doc". - Opt-in switches: --oktoask for sparse blocker-only questions, --thinking for liberal questioning, --oktoreaddocs and --oktowebsearch for grounded clarification. - Structured finalization: Produces requirement specs, ADRs, worktasks, PR descriptions, or implementation checklists with decisions, non-goals, and open questions preserved. - Use Case: You spend ten messages dumping rough ideas for a new feature; the agent captures everything quietly, then generates a complete ticket with acceptance criteria when you ask. ## Quick Start Start a braindump session for my new feature idea and just listen until I ask you to synthesize it into a ticket.

Frequently Asked Questions about ai-brain-dump

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

FAQPage Schema
How do I braindump requirements with an AI without it interrupting?▼

Start a braindump session and the agent acknowledges it will listen and hold context without asking questions, reading files, or drafting specs. It only synthesizes an artifact when you explicitly ask it to update a ticket or create a document.

How do I let the agent ask clarifying questions during a braindump?▼

Pass the --oktoask switch for sparse, blocker-only questions or --thinking for liberal questioning about any unclear detail. Both stay non-blocking so you can keep dumping context and answer when convenient.

Does listen-first mode reduce token usage in long sessions?▼

Yes, primarily by suppressing tool calls whose results get injected into context and re-billed every turn, plus keeping responses short. It does not compress conversation history, so input tokens still grow linearly with each message.

Can the agent read code or search the web during a braindump?▼

Only with explicit opt-in switches: --oktoreaddocs permits reading local code and docs, and --oktowebsearch permits web searches. Both widen grounding only and never permit modifying files or synthesizing early.

When should I not use a listen-first capture session?▼

Avoid it when you need immediate implementation or edits, since the skill forbids modifying anything until you explicitly request synthesis. It is designed for incremental context gathering, not direct task execution.