feat(agent): add Codex-native compaction paths

cli-config.yaml.example

@@ -410,6 +410,21 @@ compression:
   # Trigger compression at this % of model's context limit (default: 0.50 = 50%)
   # Lower values = more aggressive compression, higher values = compress later
   threshold: 0.50
+
+  # Existing Codex gpt-5.5 behavior: raise Hermes' compaction trigger to 85%
+  # for the ChatGPT Codex OAuth route. Set false to opt back down to threshold.
+  codex_gpt55_autoraise: true
+
+  # Codex-native compaction paths are opt-in (default: false) to preserve
+  # Hermes' existing auxiliary summarizer behavior for existing installs.
+  # When true, Codex OAuth uses Responses API compact and Codex app-server
+  # compaction uses the app-server thread compact API.
+  codex_native_compaction: false
+
+  # Codex OAuth / Responses API compaction trigger (default: 0.85 = 85%).
+  # Used only when codex_native_compaction is true. The recent tail still uses
+  # protect_last_n below.
+  codex_responses_threshold: 0.85
   
   # Fraction of the threshold to preserve as recent tail (default: 0.20 = 20%)
   # e.g. 20% of 50% threshold = 10% of total context kept as recent messages.
@@ -422,6 +437,13 @@ compression:
   # compression of older turns.
   protect_last_n: 20
 
+  # Codex app-server auto-compaction mode:
+  # Used only when codex_native_compaction is true.
+  #   native = let Codex decide when to compact its own thread (default)
+  #   hermes = let Hermes threshold trigger Codex thread/compact/start
+  #   off    = Hermes will not auto-trigger compaction; Codex may still compact natively
+  codex_app_server_auto: native
+
   # Number of non-system messages to protect at the head of the transcript, in
   # ADDITION to the system prompt (which is always implicitly protected).
   # Head messages are NEVER summarized — they survive every compression