Pipeline Plan inline

Plan saved to docs/plans/2026-03-02-ping-command.md. Here's a summary:

---

Ping Command — Implementation Plan

4 files changed (2 created, 2 modified), 13 tasks.

Design Decision

Standalone script pattern — identical to sw-hello.sh. Rejected inline-in-router because it violates every established convention and can't be tested in isolation.

Alternatives Considered

Option	Trade-offs
Standalone script (chosen)	+Consistent, +Modular, +Testable. 3 files, minimal blast radius
Inline in router	-Inconsistent, -Untestable, -Mixes concerns. Rejected

Files

Action	File
Create	scripts/sw-ping.sh
Create	scripts/sw-ping-test.sh (6 tests)
Modify	scripts/sw (add ping) case after hello))
Modify	package.json (add ping test to npm test chain)

Definition of Done

• bash scripts/sw-ping.sh → prints pong, exits 0
• bash scripts/sw-ping-test.sh → PASS: 6 FAIL: 0
• bash scripts/sw ping → prints pong
• npm test → all suites pass

Key Risk Mitigations

• No ERR trap in test file (avoids false failures on expected non-zero exits)
• ((PASS++)) pattern matches sw-hello-test.sh exactly (bash 3.2 compatible)
• Router case placed before *) default

---

Plan complete and saved to docs/plans/2026-03-02-ping-command.md. Two execution options:

1. Subagent-Driven (this session) — I dispatch a fresh subagent per task, review between tasks, fast iteration

2. Parallel Session (separate) — Open a new session with superpowers:executing-plans, batch execution with checkpoints

Which approach?