From infinite-improbability-drive
Spawn a sandboxed LLM instance to work on a delegated task
How this skill is triggered — by the user, by Claude, or both
Slash command
/infinite-improbability-drive:spawnThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Launches an isolated LLM instance in a git worktree sandbox to handle complex tasks without polluting the host LLM's context.
Launches an isolated LLM instance in a git worktree sandbox to handle complex tasks without polluting the host LLM's context.
When you need to delegate a task to an isolated LLM:
/spawn "fix the authentication bug in src/auth.rs"
Or with mode override:
/spawn --passthrough "simple typo fix"
/spawn --aisp "complex refactoring task"
Settings in .infinite-probability/improbability-drive.toml:
[spawn]
mode = "aisp"
recovery_strategy = "moderate"
idle_timeout = 120
total_timeout = 1800
default_llm = "claude-code"
$HOME or config files--dangerously-skip-permissions is never allowed$PATH for commandsnpx claudepluginhub epiphytic/infinite-improbability-drive-v0Delegates coding tasks (debug, implement, refactor) to OpenAI Codex CLI via codex exec, skipping the Node companion runtime for faster execution. Codex writes code; Claude verifies.
Use when the user asks to run Codex CLI (codex exec, codex resume) or references OpenAI Codex for code analysis, refactoring, or automated editing
Searches MemPalace before answering questions about past work, people, projects, or prior decisions. Returns verbatim stored content instead of guessing from model memory.