docs(testing): record the setup-race timeout rule

Add a second rule to Review-Proven Test Standards: a wall-clock timeout
that gates test setup is a race budget, not a tuning knob. Size it
against measured worst-case latency on a loaded machine rather than the
idle case, record the measurement at the call site so it is not trimmed
back toward the latency it exists to clear, and make a lost setup race
say so rather than surfacing as an opaque parse error.

Generalizes the flake fixed in a41fcb3, where the escaped-writer probe
in bounded_command.rs budgeted 300ms for a descendant whose spawn-to-ready
latency measured ~30ms idle but p95 233ms / max 498ms under a loaded lane.

Also correct the section intro, which read "this is the test-quality rule"
in the singular and attributed the section wholesale to the 25-PR review
mining. This rule came from a fixed flake, not a review thread, so the
intro now separates the mined rule from later additions instead of
extending that provenance to them.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01GecRn4WhoM35gz9fRY7Y1A
Signed-off-by: mraad <mraad23@gmail.com>

Author: mraad

TESTING.md

@@ -18,9 +18,9 @@ cargo test -p buzz-test-client -- --ignored
 
 ### Review-Proven Test Standards
 
-Mined from the last 25 PRs' review threads (see Review-Proven Rules in
-[AGENTS.md](AGENTS.md)); this is the test-quality rule reviewers litigated
-most:
+The first rule below is mined from the last 25 PRs' review threads (see
+Review-Proven Rules in [AGENTS.md](AGENTS.md)) — the test-quality rule
+reviewers litigated most. The rest come from defects fixed since.
 
 **Regression tests must bind the production seam and be falsifiable.**
 A guard whose removal doesn't fail any test protects nothing — mutations
@@ -31,6 +31,20 @@ combination space (PR #6807). Scope Playwright locators — unscoped
 `getByText` in a required smoke test is a strict-mode flake (PR #6980).
 (PRs #6807, #6980, #6996, #7013)
 
+**A wall-clock timeout that gates test setup is a race budget, not a tuning
+knob.** When a fixture has to win a race against a deadline — spawning a
+helper, escaping a process group, writing a handshake file — that deadline
+bounds the *setup*, not just the behavior under test. Size it against measured
+worst-case latency on a loaded machine, never the idle case. The escaped-writer
+probe in `desktop/src-tauri/src/managed_agents/discovery/bounded_command.rs`
+budgeted 300ms for a descendant whose spawn-to-ready latency measured ~30ms
+idle but p95 233ms / max 498ms under a loaded lane, so it flaked wherever
+contention was CI-shaped. Record the measurement at the call site so the budget
+is not later trimmed back toward the latency it exists to clear. When setup does
+lose the race, fail with a message saying so — a lost setup race that surfaces
+as an opaque parse error or an assertion on missing state reads as a product
+bug and costs the debugging time twice.
+
 ---
 
 ## Live Local Relay