fix(minimax-m3-max, fable-low-effort): H-SECRET-4 rejected MAX_THINKING_TOKENS in env

The validator pattern-matches *TOKEN* as credential-shaped and rejected
both contexts, failing every task on the new harnesses. The env var name
is fixed by Claude Code, so deliver the thinking budget via the
settings.json env map in init.sh instead (jq-merge preserves apiKeyHelper
from the minimax/oauth layer).

Claude-Session: https://claude.ai/code/session_019tJk7P8tZzJ24PvtgoGhLN

harnesses/contexts/minimax-m3-max/v1/harness.yaml

@@ -8,12 +8,15 @@ provides: []
 # Model pin + effort for the MiniMax Anthropic-compatible proxy.
 # - ANTHROPIC_MODEL forces every main-loop request to M3 (1M context variant)
 #   instead of the proxy's default Claude-name mapping.
-# - MAX_THINKING_TOKENS maxes the extended-thinking budget; the MiniMax
-#   /anthropic layer translates thinking budget to M3 reasoning effort.
+# - The max thinking budget (M3 reasoning effort) is set by init.sh via
+#   settings.json — the env var name trips H-SECRET-4's credential pattern
+#   so it cannot live in this env: block.
 # - CLAUDE_CODE_AUTO_COMPACT_WINDOW matches M3's real 1M window (the proxy's
 #   model metadata under-reports 200K, which triggers premature compaction).
 env:
   ANTHROPIC_MODEL: "MiniMax-M3[1m]"
   ANTHROPIC_SMALL_FAST_MODEL: "MiniMax-M3"
-  MAX_THINKING_TOKENS: "32000"
   CLAUDE_CODE_AUTO_COMPACT_WINDOW: "1000000"
+
+scripts:
+  init: ./init.sh