Skip to content

Thinking Out Loud

Turns long rambling or voice-dictated thoughts into an audited brief you approve before any work starts.

Stars
143,726
Licence
Apache-2.0
Files
3 files

What it does

When you ramble or dictate a long, messy stream of thoughts, the agent is told to do nothing (no edits, code, or plans) until it first writes a short structured summary called an echo. The echo lists your mission, locked decisions and constraints, open questions, reversals and parked tangents, and separately lists anything the agent guessed or inferred so you can correct those first. Only after you approve does it proceed, and it offers to save the brief rather than saving it on its own.

Great for

  • People who dictate long voice messages and want the agent to confirm understanding before acting
  • Anyone who changes their mind mid-message and needs those reversals surfaced rather than silently resolved
  • Users who want to be interviewed about a fuzzy, half-formed idea before building anything

Skip it if

  • Short, clear requests, where the echo step adds delay without benefit
  • Requests for a verbatim transcript or cleanup of dictation that must keep exact wording, because the skill tells the agent to resolve dictation artifacts silently
  • Pasted specs or documents that are already structured, which the skill says it should not be used for

Try asking

  • “Let me think out loud for a bit about the new onboarding flow before we build anything.”
  • “Switching to voice, sorry for typos: we need the report to cover the usual regions, actually no, scrap that, just the EU.”
  • “I have a fuzzy idea for a side project, interview me about it before you start.”
  • “Capture mode: I'm going to ramble across a few messages, say done when I'm finished.”
Step-by-step guide: how to use Thinking Out LoudInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 3 original files, the Apache-2.0 licence and an attribution note.

    thinking-out-loud.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip thinking-out-loud.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip thinking-out-loud.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Say explicitly when you want the echo, and when you want the interview, so the agent does not jump ahead.
  • Read the My additions section first and correct guesses like vague quantifiers before approving the brief.
  • For multi-message rambles, invoke capture mode and end with done to trigger the echo.

What's inside

  • README.md
  • references/echo-format.md
  • SKILL.md

Redistributed unmodified from Shubhamsaboo/awesome-llm-apps under the Apache-2.0 licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.