skills/research/SKILL.md
Research a topic and distill into reference docs. Use 'deep' for parallel multi-source research.
npx skillsauth add MansoorMajeed/Clawd researchInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
Your goal is to gather information and distill it into clean reference material — NOT to write code.
Persist to .scratch/research/ when:
Keep in context only when:
Write to .scratch/research/YYYY-MM-DD-<slug>.md. The slug should be descriptive enough to scan in ls.
Keep the format simple:
# <Title>
Date: YYYY-MM-DD
Context: <one-line — what feature/change this supports>
<distilled content — findings, code examples, source links, open questions>
No elaborate frontmatter. The value is in the content.
If .scratch/research/ exists and is non-empty, scan filenames before starting. Prior research may already answer part of the question — read relevant files to avoid re-deriving known information.
After research is complete, ask: should any of this graduate to llm-context/ or docs/?
.scratch/research/ are scratch; committed docs are curated.scratch/research/ files when they've served their purpose.scratch/ should be gitignored globally so it works in any repo without per-repo config:
echo '.scratch/' >> ~/.gitignore_global
git config --global core.excludesFile ~/.gitignore_global
tools
Allows to interact with web pages by performing actions such as clicking buttons, filling out forms, and navigating links. It works by remote controlling Google Chrome or Chromium browsers using the Chrome DevTools Protocol (CDP). When Claude needs to browse the web, it can use this skill to do so.
development
Update llm-context/ and CLAUDE.md to reflect recent code changes.
tools
Remote control tmux sessions for interactive CLIs (python, gdb, etc.) by sending keystrokes and scraping pane output.
development
Fetch a URL or convert a local file (PDF/DOCX/HTML/etc.) into Markdown using `uvx markitdown`, optionally it can summarize