Skip to content
Get packs

Structured Deep Reasoning

Walks the agent through a 10-stage reasoning loop for hard decisions, reviews and trade-offs

Source Download skill
Repository
AgriciDaniel/claude-obsidian
Stars
16,065
Licence
MIT
Files
1 file

What it does

When you ask the agent to think a hard decision through, this skill makes it follow a ten-step review process: check the evidence, question its own biases, look for hidden connections and risks, argue against its preferred answer, and end with a recommendation plus a plan to revisit it. It separates observed facts from guesses and unknowns, and it is explicitly read-only, so it reasons and recommends rather than changing files or publishing anything.

Great for

  • Engineers preparing architecture reviews or postmortems who want a structured challenge to their reasoning
  • People facing a consequential or irreversible decision who want assumptions and counterarguments surfaced
  • Anyone who wants a deliberate devil's-advocate pass on a plan before committing

Skip it if

  • Simple, reversible lookups or typo fixes, which the skill itself says should skip the formal loop; it adds overhead without value
  • Situations needing a guaranteed, deterministic check or automated repair, since the skill is a reasoning review rather than a health-check tool
  • Treating its output as verified fact: it is a reasoning framework and its conclusions still depend on the evidence the user supplies

Try asking

  • “Think this through: should we split our monolith into services or keep it as one deployable?”
  • “Do a deep think on whether to migrate our database this quarter, including the risks.”
  • “Run an architecture review of this API design and challenge my assumptions.”
  • “Write a postmortem analysis of last night's outage with counterarguments to our leading explanation.”
Step-by-step guide: how to use Structured Deep ReasoningInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 1 original files, the MIT licence and an attribution note.

    think.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip think.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip think.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

  • Give the agent concrete evidence (logs, documents, metrics) so the Observed and Unknown sections are grounded rather than guessed.
  • State the decision, the success condition and what would be irreversible up front so the LISTEN and ACCEPT stages have something to test against.
  • If the review recommends a change, explicitly authorise the follow-up action in a separate step, since this skill does not edit files.

What's inside

  • SKILL.md

Redistributed unmodified from AgriciDaniel/claude-obsidian under the MIT 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.