Skip to content
Back to search

claude-skills-minimal

A small set of Claude Code skills for research work (review, reports, long-running jobs)

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today
Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/lucas-schwengber--claude-skills-minimal/badge.svg)](/listings/lucas-schwengber--claude-skills-minimal)

HTML

<a href="/listings/lucas-schwengber--claude-skills-minimal"><img src="/api/listings/lucas-schwengber--claude-skills-minimal/badge.svg" alt="findsafeskills safety rating" /></a>

Install

i

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

git clone https://github.com/Lucas-Schwengber/claude-skills-minimal.git ~/dev/claude-skills-minimal
cd ~/dev/claude-skills-minimal
./install.sh

`install.sh` symlinks each skill folder into `~/.claude/skills/` (override with `CLAUDE_SKILLS_DIR`). It also links the hook scripts and registers them in `~/.claude/settings.json` (see Hooks), and sets this clone's git `core.hooksPath` to `hooks/`, so a later `git pull` that adds or removes a skill relinks automatically. Re-run it by hand any time. Or just copy a skill folder into `~/.claude/skills/`. Each skill is then a slash command named after its folder, e.g. `/parallel-review`.

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

README.md · README · 58 lines
  1. # claude-skills-minimal
  2. A small set of [Claude Code](https://code.claude.com/docs/en/skills) skills I use for
  3. research work (writing, reviewing, and long-running experiments). This is a trimmed copy
  4. of a personal repo, shared as is and **not actively maintained**.
  5. ## Install
  6. ```bash
  7. git clone https://github.com/Lucas-Schwengber/claude-skills-minimal.git ~/dev/claude-skills-minimal
  8. cd ~/dev/claude-skills-minimal
  9. ./install.sh
  10. ```
  11. `install.sh` symlinks each skill folder into `~/.claude/skills/` (override with
  12. `CLAUDE_SKILLS_DIR`). It also links the hook scripts and registers them in
  13. `~/.claude/settings.json` (see Hooks), and sets this clone's git `core.hooksPath` to
  14. `hooks/`, so a later `git pull` that adds or removes a skill relinks automatically.
  15. Re-run it by hand any time. Or just copy a skill folder into `~/.claude/skills/`.
  16. Each skill is then a slash command named after its folder, e.g. `/parallel-review`.
  17. ## Skills
  18. | Skill | What it does |
  19. |---|---|
  20. | `parallel-review` | Reviews a slide deck, paper draft or report with four parallel read-only reviewers (language, consistency, math and argument, literature), verifies the findings, and writes one checkbox-tracked review file. Later applies the fixes you approve. |
  21. | `write-report` | Conventions for Markdown reports and notes: YAML frontmatter, `(H)`/`(C)` author tags (human / Claude) on sections, experiment write-up structure, notation, length. Includes `check_report.py` to lint a report. |
  22. | `interrogate-me` | Claude interviews you about a plan or design, one decision at a time, with a recommended answer for each. |
  23. | `parallel-work` | How to split a task across sub-agents: when to delegate, how to brief them, how to check what they return. |
  24. | `supervised-implement` | Implement a plan together, one reviewed edit at a time, with proposals shown in chat before they reach a file. |
  25. | `docstring` | NumPy-style Python docstrings and section-divider comments. |
  26. | `digest-make` | Copies a report with its sections blanked, so you can fill in your own comments and questions and have Claude answer them. Pairs with `write-report`'s tags. |
  27. | `convert-notes` | Transcribes handwritten notes in a PDF to Markdown with LaTeX math. |
  28. | `resume-later` | Wraps up a session before you leave: finds work that would die with it, relaunches or records it, and writes a handoff note with ready-made checks. |
  29. | `clean-sessions` | Finds dangling Claude Code processes (stuck, orphaned, long idle) and asks which to end. |
  30. | `resources-summary` | One table of the jobs and processes behind a session: node, cores, memory, GPUs, elapsed, predicted finish. |
  31. ## Hooks
  32. `global/hooks/` holds Claude Code hook scripts; `global/hooks.json` has the shape of
  33. settings.json's `"hooks"` object. `install.sh` links each script into `~/.claude/hooks/`
  34. and merges the manifest into `~/.claude/settings.json` (needs `jq`; it backs up an
  35. existing file first, never adds an entry twice, and leaves your own hooks alone). Restart
  36. running sessions after the first install, since Claude Code reads hooks at startup.
  37. - `fork-rename.sh` (`UserPromptSubmit`): when a Claude **desktop app** session's title is
  38. still the default `<parent title> (fork)` / `(fork N)`, it tells Claude to rename it to
  39. `↳ <what this fork is for> (fork N)`. It reads `~/.claude/sessions/*.json`, which only
  40. the desktop app writes; elsewhere it does nothing.
  41. ## Notes
  42. - `resume-later` and `resources-summary` assume a **Slurm** cluster and a Linux login
  43. machine, and `resume-later` needs `gawk`. `clean-sessions` is Linux-only.
  44. - `interrogate-me` is a lightly adapted version of an idea from
  45. [mattpocock/skills](https://github.com/mattpocock/skills).
  46. - Skills can run shell commands and pre-approve tools. Read a skill before installing it.
  47. - Skills are plain Markdown. Edit them freely to match your own conventions.

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.