Manage AFK mode — agent operates autonomously while you're away. Verbs is `on` (default) / `off` / `status`. Bare /afk = enable. /afk "task" = enable with task context. /afk on/off/status [task] for explicit verb. As of v0.3.0 this is one consolidated skill (was three).
Break down any concept, system, code, or decision using the Feynman-technical hybrid formula — plain-language mechanism → evidence/numbers → actionable implication. Use whenever the user asks to explain, break down, simplify, or understand something. Trigger on "feynman this", "ELI5", "break this down", "help me understand", "explain like I'm five", "walk me through", "what does X actually mean", "why does X work", or any request to make something clearer. Also use proactively when explaining architecture decisions, engineering trade-offs, research findings, or anything where both accessibility and rigor matter. When in doubt, use this skill — clarity is never wrong.
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
Reusable agent skills for Claude Code, Codex, and OpenCode. Install once, use across surfaces. Each skill is a self-contained directory with a SKILL.md, optional supporting scripts, and its own README.md.
| Skill | What it does | Audience |
|---|---|---|
afk | Step away from your keyboard and have your agent keep working autonomously. Subcommands: /afk (or /afk on) enables, /afk off disables, /afk status inspects. | Claude Code power users running long sessions |
Add this repo as a plugin marketplace, then install the bundle:
/plugin marketplace add ming1in/skills
/plugin install ming-skills@ming-skills
As of v0.4.0, lifecycle hooks ship inside the plugin (declared in .claude-plugin/plugin.json hooks field) and activate automatically when the plugin is enabled. No separate install step is required — the /plugin install above is the entire setup.
Migrating from v0.3.x? Earlier versions required
bash skills/afk/install.shto register hooks in your~/.claude/settings.json. Those entries are now redundant (and will fire alongside the bundled hooks, doubling every event). Clean them up before upgrading: any line in~/.claude/settings.jsonhookscontaining# afk-skillshould be removed. One-liner:python3 -c "import json,os; p=os.path.expanduser('~/.claude/settings.json'); s=json.load(open(p)); s['hooks']={e:[g for g in groups if not any('afk-skill' in h.get('command','') for h in g.get('hooks',[]))] for e,groups in s.get('hooks',{}).items()}; s['hooks']={k:v for k,v in s['hooks'].items() if v}; json.dump(s,open(p,'w'),indent=2)"
INSTALLER="$HOME/.codex/skills/.system/skill-installer/scripts/install-skill-from-github.py"
python3 "$INSTALLER" --repo ming1in/skills --path skills/afk
As of v0.3.0 there's only one skill (afk) — earlier versions had three (afk, afk-off, afk-status) and you'd pass them all to one --path invocation. The repeatable --path mechanism still works for future skills.
Restart Codex after installing. The installer drops files into $CODEX_HOME/skills/ (default ~/.codex/skills/). It aborts if a destination already exists — to update, delete the destination directory and re-run.
Note: AFK's autonomous Stop-hook behavior currently uses Claude-Code-specific env substitutions (
${CLAUDE_SKILL_DIR},${CLAUDE_SESSION_ID}). Codex has compatibleSessionStartandStophooks (theStophook supports the samedecision: "block"continuation pattern), so the AFK-on-Codex adapter is a small port — not blocked on missing platform features. Until that adapter ships, installing under Codex makes the skill source available for review.
Install path TBD — see .opencode/INSTALL.md.
If you've cloned this repo (or use it as a git submodule like big-one does at submodules/skills), point Claude Code at the local checkout instead of GitHub:
/plugin marketplace add ./submodules/skills
/plugin install ming-skills@ming-skills
For Codex, symlink the skill directory into $CODEX_HOME/skills/:
mkdir -p "${CODEX_HOME:-$HOME/.codex}/skills"
ln -s "$PWD/skills/afk" "${CODEX_HOME:-$HOME/.codex}/skills/afk"
# Plugin registered, hooks installed, /afk available
# Start working with Claude on something long-running:
"Help me refactor the user-auth module"
…work, work…
# Step away — AFK keeps the agent going on the in-flight work:
/afk
# (close laptop, get coffee)
# Check what happened:
/afk status
# Done?
/afk off
The agent keeps working through Stop events until it self-marks task_status: "done" (work complete) or task_status: "blocked" (hard approval gate hit), or you run /afk off. A 50-iteration cap prevents runaway loops.
.
├── .claude-plugin/ # Claude Code marketplace + plugin manifests
├── .codex-plugin/ # Codex plugin manifest
├── .opencode/ # OpenCode install notes (placeholder)
├── .github/workflows/ # CI: shellcheck + JSON validation + AFK smoke test
├── skills/
│ └── afk/ # /afk slash command (verbs: on/off/status) + scripts
├── CHANGELOG.md
├── LICENSE # MIT
└── README.md
The afk directory has a SKILL.md (the slash-command body that dispatches on verb), a README.md (human-facing docs), and bundled scripts/. Lifecycle hooks (Stop, SessionEnd, SessionStart) are declared in .claude-plugin/plugin.json and activate automatically with the plugin.
npx claudepluginhub ming1in/skills --plugin skillsPersistent file-based planning for AI coding agents. Crash-proof markdown plans (task_plan.md, findings.md, progress.md) that survive context loss and /clear, with an opt-in completion gate and multi-agent shared state. Manus-style. Works with Claude Code, Codex CLI, Cursor, Kiro, OpenCode and 60+ agents via the SKILL.md standard. Includes Arabic, German, Spanish, and Chinese (Simplified and Traditional).
Claude harness - A harness for solo developers (Vibecoders) to handle full-cycle contract development.
Core skills library for Claude Code: TDD, debugging, collaboration patterns, and proven techniques
Harness-native ECC operator layer - 67 agents, 271 skills, 92 legacy command shims, reusable hooks, rules, selective install profiles, and production-ready workflows for Claude Code, Codex, OpenCode, Cursor, and related agent harnesses
v9.44.1 — Patch release for Gemini environment/version detection and qwen auth gating. Run /octo:setup.
Superpowers Plus core skills library for Claude Code: planning, execution routing, TDD, debugging, and collaboration workflows