housekeeping — for Claude Code housekeeping, cc-disco, community, for Claude Code, ide skills, bash git status git diff --stat, gitignore, git push, gardening, workspace

v1.0.0

Sobre este Skill

Cenario recomendado: Ideal for AI agents that need step 1: git hygiene. Resumo localizado: # Housekeeping This skill is repo gardening: tidy the workspace, consolidate knowledge, and leave things cleaner than you found them. This AI agent skill supports Claude Code, Cursor, and Windsurf workflows.

Recursos

Step 1: Git hygiene
Untracked files: decide committed vs. gitignored.
Default: commit unless clearly transient (tmp files, screenshots mid-session, build artifacts
If unsure: commit. Git history is cheap, lost knowledge is not.
Staged/modified files: commit what makes sense as a logical unit. Don't batch unrelated changes

# Tópicos principais

mikesol mikesol
[2]
[0]
Atualizado: 3/20/2026

Skill Overview

Start with fit, limitations, and setup before diving into the repository.

Cenario recomendado: Ideal for AI agents that need step 1: git hygiene. Resumo localizado: # Housekeeping This skill is repo gardening: tidy the workspace, consolidate knowledge, and leave things cleaner than you found them. This AI agent skill supports Claude Code, Cursor, and Windsurf workflows.

Por que usar essa habilidade

Recomendacao: housekeeping helps agents step 1: git hygiene. Housekeeping This skill is repo gardening: tidy the workspace, consolidate knowledge, and leave things cleaner than you found them. This AI agent skill

Melhor para

Cenario recomendado: Ideal for AI agents that need step 1: git hygiene.

Casos de Uso Práticos for housekeeping

Caso de uso: Step 1: Git hygiene
Caso de uso: Untracked files: decide committed vs. gitignored
Caso de uso: Default: commit unless clearly transient (tmp files, screenshots mid-session, build artifacts

! Segurança e Limitações

  • Limitacao: Staged/modified files: commit what makes sense as a logical unit. Don't batch unrelated changes into one commit.
  • Limitacao: Don't batch unrelated changes into one commit
  • Limitacao: Requires repository-specific context from the skill documentation

About The Source

The section below is adapted from the upstream repository. Use it as supporting material alongside the fit, use-case, and installation summary on this page.

Demo Labs

Browser Sandbox Environment

⚡️ Ready to unleash?

Experience this Agent in a zero-setup browser environment powered by WebContainers. No installation required.

Boot Container Sandbox

FAQ e etapas de instalação

These questions and steps mirror the structured data on this page for better search understanding.

? Perguntas frequentes

O que é housekeeping?

Cenario recomendado: Ideal for AI agents that need step 1: git hygiene. Resumo localizado: # Housekeeping This skill is repo gardening: tidy the workspace, consolidate knowledge, and leave things cleaner than you found them. This AI agent skill supports Claude Code, Cursor, and Windsurf workflows.

Como instalar housekeeping?

Execute o comando: npx killer-skills add mikesol/cc-disco. Ele funciona com Cursor, Windsurf, VS Code, Claude Code e mais de 19 outros IDEs.

Quais são os casos de uso de housekeeping?

Os principais casos de uso incluem: Caso de uso: Step 1: Git hygiene, Caso de uso: Untracked files: decide committed vs. gitignored, Caso de uso: Default: commit unless clearly transient (tmp files, screenshots mid-session, build artifacts.

Quais IDEs são compatíveis com housekeeping?

Esta skill é compatível com Cursor, Windsurf, VS Code, Trae, Claude Code, OpenClaw, Aider, Codex, OpenCode, Goose, Cline, Roo Code, Kiro, Augment Code, Continue, GitHub Copilot, Sourcegraph Cody, and Amazon Q Developer. Use a CLI do Killer-Skills para uma instalação unificada.

housekeeping tem limitações?

Limitacao: Staged/modified files: commit what makes sense as a logical unit. Don't batch unrelated changes into one commit.. Limitacao: Don't batch unrelated changes into one commit. Limitacao: Requires repository-specific context from the skill documentation.

Como instalar este skill

  1. 1. Abra o terminal

    Abra o terminal ou linha de comando no diretório do projeto.

  2. 2. Execute o comando de instalação

    Execute: npx killer-skills add mikesol/cc-disco. A CLI detectará sua IDE ou agente automaticamente e configurará a skill.

  3. 3. Comece a usar o skill

    O skill já está ativo. Seu agente de IA pode usar housekeeping imediatamente no projeto atual.

! Source Notes

This page is still useful for installation and source reference. Before using it, compare the fit, limitations, and upstream repository notes above.

Upstream Repository Material

The section below is adapted from the upstream repository. Use it as supporting material alongside the fit, use-case, and installation summary on this page.

Upstream Source

housekeeping

# Housekeeping This skill is repo gardening: tidy the workspace, consolidate knowledge, and leave things cleaner than you found them. This AI agent skill

SKILL.md
Readonly
Upstream Repository Material
The section below is adapted from the upstream repository. Use it as supporting material alongside the fit, use-case, and installation summary on this page.
Upstream Source

Housekeeping

This skill is repo gardening: tidy the workspace, consolidate knowledge, and leave things cleaner than you found them. Work systematically through the steps below, committing progress as you go.


Step 1: Git hygiene

bash
1git status 2git diff --stat
  • Untracked files: decide committed vs. gitignored.
    • Default: commit unless clearly transient (tmp files, screenshots mid-session, build artifacts, secrets).
    • If unsure: commit. Git history is cheap, lost knowledge is not.
  • Staged/modified files: commit what makes sense as a logical unit. Don't batch unrelated changes into one commit.
  • gitignore: if anything clearly shouldn't be tracked (logs, large caches, .env), add it to .gitignore.
  • Push: after committing, git push.

Step 2: Upstream sync

cc-disco forks can optionally maintain a clone of the upstream repo for syncing improvements and contributing back. Check CLAUDE.md for a configured upstream clone path.

If an upstream clone exists:

bash
1cd <upstream-clone-path> 2git fetch origin 3git log HEAD..origin/main --oneline # what's new upstream?

If there are upstream changes:

  1. Review them — are they relevant? Do they conflict with local customizations?
  2. Pull them into the live fork:
    bash
    1cd <live-repo-path> 2git remote add upstream <upstream-url> # if not already added 3git fetch upstream 4git merge upstream/main
  3. Resolve conflicts carefully — local customizations (CLAUDE.md, skills, data/) take precedence over upstream defaults unless upstream has a clearly better approach.
  4. Commit the merge and push.

Also consider: if local improvements were made during this housekeeping session that would benefit other cc-disco instances, mention them as upstream contribution candidates.

If no upstream clone is configured: Ask the user: "Would you like to set up an upstream clone? It lets you pull in improvements from the cc-disco community and contribute back. I can clone it to a scratch directory and configure it in CLAUDE.md."

If they say yes, clone the upstream repo to a suitable location (e.g. ~/scratch/cc-disco-upstream), add the path and upstream URL to CLAUDE.md, and proceed with the sync.


Step 3: Recent channel review

Look at Discord messages from the past ~24 hours across all channels to spot undocumented knowledge.

Use the Discord API (bot token and guild ID from CLAUDE.md / 1Password). Fetch recent messages from each channel in the configured guild.

For each message thread or conversation, ask:

  • Is this a recurring task or workflow? → candidate for a new skill
  • Is this a fact about the world, a person, a service, or a codebase? → candidate for CLAUDE.md or a knowledge file in the repo
  • Did something break or get fixed? → candidate for a gotcha note in the relevant skill

Document anything worth keeping. Update CLAUDE.md or the relevant SKILL.md. If something clearly warrants its own skill, draft it.


Step 4: Organizational reflection

Step back and look at the repo as a whole. The goal is to notice where organic growth has created mess — and fix it.

  • Knowledge consolidation: is related information scattered across multiple files or skills? If so, merge or cross-reference.
  • Skills: are any skills redundant, overlapping, or half-finished? Could two be merged into one tighter skill? Is there a skill that used to make sense but no longer reflects how things work?
  • Structure: does the directory layout still make sense? Are there files or folders that feel out of place?
  • Pruning: is there anything that simply shouldn't exist anymore — dead code, abandoned ideas, outdated knowledge files? Remove it.
  • Naming: are things named clearly and consistently?

This step is judgment-driven, not checklist-driven. The question is: if someone came to this repo fresh today, would the layout feel coherent? If not, fix what you can and flag what needs a bigger conversation.


Step 5: CLAUDE.md audit

Read CLAUDE.md from top to bottom. Check for:

  • Stale references: file paths, URLs, credential IDs, channel IDs that may have changed
  • Outdated facts: skills or infrastructure that no longer exist or have changed
  • Bloat: sections that have grown accretive and incoherent — rewrite them tightly
  • Missing entries: new skills, tools, or facts that should be indexed but aren't

Edit CLAUDE.md in place. Keep it light. It is an index and orientation guide, not a full manual.


Step 6: Skills audit

For each skill in .claude/skills/:

  • Does it still work? Compare the SKILL.md description against what you know about the current state of the tool/service.
  • Are credentials still accurate? Check 1Password item IDs and field names referenced.
  • Are there gotchas missing? If a skill has caused problems, document them.
  • Is the skill redundant? If two skills overlap heavily, consider merging them.

If a skill is clearly obsolete, remove it and update CLAUDE.md.


Step 7: Working files check

Look for any non-source files in the repo — state files, caches, logs, screenshots, build artifacts, downloaded content — and make sure each is handled correctly:

  • Should it be committed? If it's meaningful persistent state (e.g. deduplication state, config), commit it.
  • Should it be gitignored? If it's ephemeral or large, add it to .gitignore.
  • Should it be deleted? If it has no ongoing purpose, remove it.

There's no assumed directory structure here — scan the whole repo.


Step 8: Final commit and report

Commit any remaining changes:

bash
1git add -A 2git commit -m "Housekeeping YYYY-MM-DD" 3git push

Post a brief summary to the configured general/status channel (see CLAUDE.md) with:

  • What was tidied (files committed, docs updated, etc.)
  • Any new skills or knowledge added
  • Upstream sync status
  • Anything flagged that needs attention (conflicts you couldn't resolve, decisions you deferred, etc.)

Habilidades Relacionadas

Looking for an alternative to housekeeping or another community skill for your workflow? Explore these related open-source skills.

Ver tudo

openclaw-release-maintainer

Logo of openclaw
openclaw

Resumo localizado: 🦞 # OpenClaw Release Maintainer Use this skill for release and publish-time workflow. It covers ai, assistant, crustacean workflows. This AI agent skill supports Claude Code, Cursor, and Windsurf workflows.

widget-generator

Logo of f
f

Resumo localizado: Generate customizable widget plugins for the prompts.chat feed system # Widget Generator Skill This skill guides creation of widget plugins for prompts.chat. It covers ai, artificial-intelligence, awesome-list workflows. This AI agent skill supports Claude Code, Cursor, and

flags

Logo of vercel
vercel

Resumo localizado: The React Framework # Feature Flags Use this skill when adding or changing framework feature flags in Next.js internals. It covers blog, browser, compiler workflows. This AI agent skill supports Claude Code, Cursor, and Windsurf workflows.

138.4k
0
Navegador

pr-review

Logo of pytorch
pytorch

Resumo localizado: Usage Modes No Argument If the user invokes /pr-review with no arguments, do not perform a review. It covers autograd, deep-learning, gpu workflows. This AI agent skill supports Claude Code, Cursor, and Windsurf workflows.

98.6k
0
Desenvolvedor