Parallel Agents prompt

Add the following instruction to my global CLAUDE.md (the user-level file,
e.g. ~/.claude/CLAUDE.md - create the file if it doesn't exist).

Safety first: read my current global CLAUDE.md before changing anything. If it
already contains this instruction, or an equivalent rule about parallel tool
calls and subagent delegation, tell me what you found, change nothing, and
stop. If it clashes with or contradicts any instruction I already have, don't
append either - point out the clash, explain the trade-off, and ask me how I
want to handle it. Never remove or rewrite my existing content - only append.

Append this section exactly:

## Parallel Agents
Distinguish two kinds of parallelism. Parallel tool calls (multiple
Read/Grep/Bash in one message) are cheap - do them whenever the calls are
independent. This should be the default. Spawning subagents or agent teams is
heavier - only worth it when each agent has substantial, independent work.
Don't manufacture parallelism for small tasks, and don't spawn an agent to
re-read files already loaded in the current session.

When you're done, show me the section you added and confirm where it lives.