Fix Design Decision Gate false failures caused by the LLM invocation cap

[Copilot]

Three independent monitors flagged the merge-blocking Design Decision Gate 🏗️ workflow failing at a high rate while still burning real token spend — a Turns=0-style crash signature that turned out to be something else entirely.

Root cause

max-turns doubles as the AWF API proxy's hard per-run LLM invocation cap. This gate's ADR-review task typically needs ~20-22 turns to complete, but max-turns was set to 20 — so it reliably hit 429 Maximum LLM invocations exceeded (20/20). Worse, the agent had often already written a valid safe-output (e.g. an add_comment with the completed ADR review) before hitting the cap, but every harness (claude/codex/copilot) treated the cap error as an unconditional hard failure and discarded that completed work.

Changes

• Increased max-turns from 20 → 30 in design-decision-gate.md (recompiled lock file) to give headroom above the observed completion point.
• claude_harness.cjs, codex_harness.cjs, copilot_harness.cjs: when the invocation cap is saturated but safe-outputs already contain a terminal/expected result, exit 0 instead of failing — mirroring the existing suppression pattern already used for "numerous permission-denied" and "partial execution" cases.
• Added harness tests covering the new suppression behavior for all three engines.

// Before: cap saturation was always fatal, even with completed work
if (nonRetryableGuard.maxRunsExceeded) return { action: "stop" }; // exit 1

// After: suppress the false-red when safe-outputs already hold the result
if (nonRetryableGuard.maxRunsExceeded && hasExpectedSafeOutputs(safeOutputsPath)) {
  return { action: "stop", exitCode: 0 };
}

• Fixes [deep-report] Investigate Design Decision Gate merge-blocking failure/cost hotspot (cross-verified by 3 monitors) #52736

[github-actions]

PR Triage

Category: bug | Risk: high | Score: 82/100

• Impact: 35/50 (fixes merge-blocking Design Decision Gate false failures burning real token spend across the org)
• Urgency: 30/30 (flagged by three independent monitors — actively impacting CI reliability)
• Quality: 17/20 (no reviews yet, CI status unknown, but clear root-cause narrative)

Recommended action: fast_track
High-urgency CI reliability fix — actively blocking merges and burning token spend. Batched with reliability-fixes cluster (#52757, #52837, #52854).

Generated by 🔧 PR Triage Agent · auto · 46.6 AIC · ⌖ 2.63 AIC · ⊞ 8.1K · ◷