Add chapter: subprocesses and external streams (structured teardown)

[adamw]

What

Adds guidance for the case structured concurrency is easiest to get subtly wrong: driving a long-lived subprocess (or socket / SSE) whose output is read line by line.

• New chapter 170-subprocesses-and-external-streams.md:
  • own the work with a supervised scope; the reader's return value is the result (fork{…}.join()), not a cross-thread AtomicReference/CAS;
  • never return a live object that owns running forks to be driven later;
  • teardown: a blocking native read ignores interruption, and Ox joins forks before running releaseAfterScope finalizers — so destroy/close the resource in the scope body's finally, before the join;
  • kill the whole process tree (a launcher's forked child inherits the pipe fds; an orphan keeps the pipe open so the reader never EOFs);
  • drain/inherit stderr so an unread pipe can't stall the child.
• SKILL.md: two always-loaded design rules (decide the owning scope first; result is the fork's return value; never return a fork-owning handle; destroy blocking resources in the body finally, not releaseAfterScope; tree-kill launchers) plus the index entry.

Why

The detailed how lives in the chapter; the decision triggers live in the always-loaded SKILL.md so the agent picks the right ownership and teardown shape before writing code — the failure mode here is a shutdown-time deadlock that doesn't show up until teardown.

Grounded in Ox's documented scope-teardown semantics (interruptAllAndJoinUntilCompleted runs in the scope body's finally; releaseAfterScope finalizers run after).

🤖 Generated with Claude Code