[WIP] Add resilience testing plan

[Copilot]

• Understand the resilience testing plan in docs/resilience-testing.md
• Phase 1 (Bug-finding):
  • 38_infinite_loop.sql — B1/B2: Infinite/always-true loop cancelled via df.cancel()
  • 39_truthiness_edge_cases.sql — B3: Condition edge cases (NULL, "false", 0.0, empty string) — documents str_false/str_no are truthy (known quirk); 10s timeout with explicit fallthrough handling
  • 40_empty_dml_results.sql — B5/B6: SQL nodes returning 0 rows / DML without RETURNING
  • 41_break_outside_loop.sql — B10: df.break() at top level (no enclosing loop)
  • 42_recursive_start.sql — B11: df.start() called inside a workflow SQL node
  • 43_empty_sql.sql — C1: Empty, whitespace-only, and non-SQL text
  • 44_crafted_json.sql — C7: Bypassing DSL with raw JSON inputs
• Phase 2 (Stress):
  • 45_concurrent_instances.sql — A1: 20 concurrent instances tracked via df.instances join
  • 46_deep_nesting.sql — A2: 50-level sequential THEN chain
  • 47_rapid_cancel.sql — A7: 20 rapid start/cancel cycles tracked via df.instances join
• Phase 3 (Error edge cases):
  • 48_race_both_fail.sql — B8: RACE where both branches fail (df.race + | operator)
  • 49_join_one_fails.sql — B9: JOIN where one branch fails (df.join + & operator)
  • 50_signal_edge_cases.sql — B12/B13: Signal to non-existent/completed instances, duplicate signals

---

✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.