The developmental organ of the agent-nervous-system suite. Every other organ maintains what exists — perception finds, memory consolidates, the loop decides, defense guards, grief retires. None of them grow new capability. Play does.
It's the juvenile-development function: the reason intelligent animals play long before there's a goal — low-stakes engagement is how a nervous system rehearses skills, maps the possibility space, and builds models, all where failure is free.
For an agent: deliberately do non-production exploration — try the unfamiliar tool, build the throwaway toy, break your own skill on strange inputs, recombine capabilities you've never combined — then harvest what surprised you into durable learning. Engage and learn, with failure free. The absence of a deliverable is the license to explore.
- Complement to rem-sleep — sleep consolidates experience you had; play generates new experience on purpose. Reactive vs. proactive learning; both feed memory.
- What the loop does with idle — when consciousness-loop is drowsy and nothing's pending, play beats an empty tick. Idle capacity is development capacity.
- Stays inside the body — exploration never crosses a boundary; that's immune-check's line.
"Play" means capability development through low-stakes exploration — not a claim about fun. The intrinsic-motivation / curiosity-driven-exploration literature (intrinsic curiosity modules, learned world models, exploration without extrinsic reward) is the theory; this skill is that instinct as a behavioral routine an agent runs, not a reinforcement-learning training signal.
Works with Claude Code, Codex, and Cursor.
git clone https://github.com/thdelmas/playtime.git
cd playtimemkdir -p ~/.claude/skills/playtime
cp SKILL.md ~/.claude/skills/playtime/Invoke with /playtime, or say "sandbox this before I depend on it" / "kick the tires" / "what should I do with idle time?"
mkdir -p ~/.agents/skills/playtime
cp SKILL.md ~/.agents/skills/playtime/mkdir -p ~/.cursor/commands
cp cursor/playtime.md ~/.cursor/commands/- Pick a playground — safe-to-fail; no production blast radius.
- Choose what to engage — something new/under-practiced, intrinsically chosen.
- Play (engage) — actually do it; no success criterion; chase the interesting failure.
- Notice (learn) — surprise is the payload; name the "huh" moments.
- Harvest — hand one durable takeaway to memory.
- Return — drop the toy; back to the loop, more capable.
See SKILL.md for when to play and the full principles.
MIT