Skip to content
Get packs

Bounded Autoresearch

Runs a limited, source-cited web research loop and drafts a dossier for your vault after review.

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

What it does

When you ask for deep research on a topic, the agent plans a bounded set of questions, looks for official and primary sources (with the web only if you approve), and collects claims with citations, noting contradictions and gaps. It produces a draft dossier rather than changing your notes directly. Only after you review a proposed change does it file the dossier into your Obsidian-style knowledge vault, and any merge into existing canonical pages is a separate step you can accept, narrow, or reject.

Great for

  • People who want a sourced research summary on a topic and a record of where each claim came from
  • Obsidian vault users who want research filed into a structured wiki after human review
  • Anyone who wants research limited by explicit round, fetch, and page budgets

Skip it if

  • Quick factual questions, since the multi-round loop and review steps are slow and heavy (overkill for simple lookups)
  • Unattended or fully autonomous publishing into a knowledge base, because the skill requires explicit approval before writes and web access needs consent
  • Research on private or confidential material, because the skill warns against sending private vault text, paths or credentials to external services and the web step can leak context if not carefully scoped

Try asking

  • “Research and file a dossier on the history of solar panel efficiency, citing primary sources”
  • “Deep dive into current EU AI transparency rules, with approved domains only”
  • “Investigate competing explanations for coral reef decline and note contradictions”
  • “Build a wiki on container security best practices from official documentation”
Step-by-step guide: how to use Bounded AutoresearchInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 2 original files, the MIT licence and an attribution note.

    autoresearch.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

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

    Claude Code / Agent SDK (this project only)

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

  • Set explicit limits up front (rounds, number of sources, page count, domains) so the agent stays within budget
  • Review the inspect output carefully, including cited claims, raw captures and file paths, before approving the apply step
  • Keep the dossier and any canonical vault merge as separate approvals, and decline egress if the topic involves private material

What's inside

  • references/program.md
  • 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.