agentleFS
Sign inStart free
NotRedFoxNotRedFox/NotRedFoxs-Claude-skills/research-workspace/CLAUDE.md

NotRedFoxs-Claude-skills / research-workspace

CLAUDE.md1 starsChanged 34 days agoMITView on GitHub

This folder is only for research. Keep it separate from code projects.

The file

We split this CLAUDE.md into the rules it gives an agent

6 rules. About 0 are common, 1 is distinctive, and 5 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 5

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Write a short brief: what the user is building, constraints, search terms.

    line 13
  2. Very specific:

    Start code-scout, paper-scout and sideways-scout in parallel with the brief.

    line 14
  3. Very specific:

    Start verifier on their output in sources/.

    line 15
  4. Very specific:

    Write reports/<topic>.md from what the verifier confirmed: what was found, why it matters, ranked things to try with a first test each, and …

    line 16
  5. Very specific:

    Every source file has a link and a Verified: line before you finish.

    line 21
  6. Very specific:

    Never cite from memory. Cite only sources marked Verified: yes or Verified: corrected.

    line 23

≈ “Common” is an estimate: at least 1 in 5 of the 26,980 CLAUDE.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
CLAUDE.md

Research workspace

This folder is only for research. Keep it separate from code projects.

Layout

  • sources/<topic>/: one Markdown file per source saved by an agent: link, date read, status (opened, abstract only or unverified), and the excerpts that matter. The verifier adds a Verified: line to each. A BRIEF.md here is fine.
  • checks/: every script used to verify something. Save it here first, then run it from here.
  • reports/<topic>.md: the final write-up.
  • logs/actions.log: written automatically by a hook, one line per action. Don't edit it.
  • logs/notes.log: yours. One line per step: what you did and what you found.

How to research

  1. Write a short brief: what the user is building, constraints, search terms.
  2. Start code-scout, paper-scout and sideways-scout in parallel with the brief.
  3. Start verifier on their output in sources/.
  4. Write reports/<topic>.md from what the verifier confirmed: what was found, why it matters, ranked things to try with a first test each, and what was dropped.

Rules

Hooks enforce the first three. You'll be stopped if you skip them.

  • Any code that checks something is saved in checks/ before it runs. Inline python -c, node -e and heredocs are blocked.
  • Every source file has a link and a Verified: line before you finish.
  • A report in reports/ is written or updated after the last source changed.
  • Never cite from memory. Cite only sources marked Verified: yes or Verified: corrected.
  • Plain writing: no em dashes, no filler, real numbers.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

Reviews can't be read right now.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free