fix: prevent infinite auto-compact loop for unknown 3P models (#635) (#636)

- Raise context window fallback from 8k to 128k for unknown OpenAI-compat models.
  The 8k fallback caused effective context (8k minus output reservation) to go
  negative, making auto-compact fire on every single message.
- Add safety floor in getEffectiveContextWindowSize(): effective context is
  always at least reservedTokensForSummary + 13k buffer, ensuring the
  auto-compact threshold stays positive.
- Add missing MiniMax model entries (M2.5, M2.5-highspeed, M2.1, M2.1-highspeed)
  all at 204,800 context / 131,072 max output per MiniMax docs.
- Add tests for MiniMax variants, 128k fallback, and autoCompact floor.

Fixes #635

Co-authored-by: root <root@vm7508.lumadock.com>

src/utils/context.test.ts

@@ -107,9 +107,23 @@ test('MiniMax-M2.7 uses explicit provider-specific context and output caps', ()
   expect(getMaxOutputTokensForModel('MiniMax-M2.7')).toBe(131_072)
 })
 
-test('unknown openai-compatible models still use the conservative fallback window', () => {
+test('unknown openai-compatible models use the 128k fallback window (not 8k, see #635)', () => {
   process.env.CLAUDE_CODE_USE_OPENAI = '1'
   delete process.env.CLAUDE_CODE_MAX_OUTPUT_TOKENS
 
-  expect(getContextWindowForModel('some-unknown-3p-model')).toBe(8_000)
+  expect(getContextWindowForModel('some-unknown-3p-model')).toBe(128_000)
+})
+
+test('MiniMax-M2.5 and M2.1 use explicit provider-specific context and output caps', () => {
+  process.env.CLAUDE_CODE_USE_OPENAI = '1'
+  delete process.env.CLAUDE_CODE_MAX_OUTPUT_TOKENS
+
+  expect(getContextWindowForModel('MiniMax-M2.5')).toBe(204_800)
+  expect(getContextWindowForModel('MiniMax-M2.5-highspeed')).toBe(204_800)
+  expect(getContextWindowForModel('MiniMax-M2.1')).toBe(204_800)
+  expect(getContextWindowForModel('MiniMax-M2.1-highspeed')).toBe(204_800)
+  expect(getModelMaxOutputTokens('MiniMax-M2.5')).toEqual({
+    default: 131_072,
+    upperLimit: 131_072,
+  })
 })