Raise Evergreen daily AI guardrail

.github/workflows/evergreen.md

@@ -20,7 +20,7 @@ on:
 permissions: read-all
 
 timeout-minutes: 30
-max-daily-ai-credits: 5000
+max-daily-ai-credits: 20000
 
 network:
   allowed:

.github/workflows/shared/evergreen/quota-policy.md

@@ -5,10 +5,12 @@ Quota is per PR and per continuous application of the `evergreen` label.
 1. Quota starts when `evergreen` is applied and continues across runs while the
    label remains.
 2. The per-PR budget is **5000 AI credits** (see `repo-policy.md`).
-3. New commits do not reset quota by themselves.
-4. Empty CI-trigger commits do not count as semantic repair attempts.
-5. Cheap deterministic monitoring runs should consume little or no quota.
-6. Reapplying `evergreen` after exhaustion starts a fresh quota; keep prior
+3. The daily Evergreen workflow guardrail is **20000 AI credits** so scheduled
+   repair runs do not stop after one or two substantial PR passes.
+4. New commits do not reset quota by themselves.
+5. Empty CI-trigger commits do not count as semantic repair attempts.
+6. Cheap deterministic monitoring runs should consume little or no quota.
+7. Reapplying `evergreen` after exhaustion starts a fresh quota; keep prior
    memory.
 
 On quota exhaustion:

.github/workflows/shared/evergreen/repo-policy.md

@@ -100,6 +100,7 @@ by the maintainer. Evergreen must follow it for all repo-specific behavior.
 
 - Per-PR AI-credit budget: **5000** AI credits per continuous application of the
   `evergreen` label.
+- Daily Evergreen workflow AI-credit guardrail: **20000** AI credits.
 - Max runs: bounded by the budget; scheduled monitoring should use little quota.
 - Max repeated attempts per failure signature: stop repeating a path already
   recorded as failing in memory.