feat(compression): make protect_first_n configurable

The number of head messages preserved verbatim across context compactions
was previously hardcoded to 3 in AIAgent.__init__. Expose it as
`compression.protect_first_n` in config, matching the existing
`protect_last_n` pattern.

Motivation: users who rely on rolling compaction for long-running sessions
had the opening user/assistant exchange pinned as head forever, which
doesn't always match how they want the session framed after many
compactions. Lowering to 1 preserves the system prompt + first non-system
message; lowering to 0 preserves only the system prompt and lets the
entire first exchange age out naturally through the summary.

Semantics: `protect_first_n` counts non-system head messages protected
**in addition to** the system prompt, which is always implicitly protected
when present. Same meaning across both code paths:

  protect_first_n=0 → system prompt only (or nothing if no system message)
  protect_first_n=2 → system prompt + first 2 non-system messages (default)

This unifies the CLI path (which reads messages with the system prompt at
position 0) and the gateway path (where the gateway /compress handler
strips the system prompt before calling compress() — see
gateway/run.py L9150-9154 on the parent fork). Previously these two paths
disagreed:

  CLI path:     protect_first_n=1 → protect system prompt only
  Gateway path: protect_first_n=1 → protect first USER turn forever

In practice on long-running gateway sessions the old semantics pinned
whatever stale aside happened to be the first user message, reinserting
it into every compaction summary indefinitely.

Default chosen as 2 (not 3) so that the effective protected head count
remains 3 messages in the common case — assuming a system prompt is
present, default protection becomes system + 2 non-system = 3 total,
matching the pre-feature behaviour where `protect_first_n` was hardcoded
to protect 3 messages total. Sessions without a system prompt will see a
small behaviour change (2 protected head messages instead of 3), but this
is the rare path and the new semantics make the system-prompt-present
case the well-defined one.

Changes:

- agent/context_compressor.py: redefine protect_first_n as the count of
  non-system head messages protected beyond the implicit system-prompt
  guarantee; both paths converge. Constructor default updated to 2.
- hermes_cli/config.py: add `compression.protect_first_n` default (2),
  matching the new semantics. `show_config` label tweaked to
  'Protect first: N non-system head messages' for clarity.
- run_agent.py: read protect_first_n from config; 0 is now valid (system
  prompt is always implicitly protected).
- cli-config.yaml.example: document the new key and rationale.
- tests/agent/test_context_compressor.py: cover default, override, the
  end-to-end `protect_first_n=0` and `protect_first_n=1` behaviour,
  the no-system-prompt (gateway) path, and the new shared-semantics
  regression test.

Fixes #13751
Tested on Ubuntu 24.04.

agent/context_compressor.py

@@ -405,7 +405,7 @@ class ContextCompressor(ContextEngine):
         self,
         model: str,
         threshold_percent: float = 0.50,
-        protect_first_n: int = 3,
+        protect_first_n: int = 2,
         protect_last_n: int = 20,
         summary_target_ratio: float = 0.20,
         quiet_mode: bool = False,
@@ -1185,6 +1185,26 @@ The user has requested that this compaction PRIORITISE preserving all informatio
             idx += 1
         return idx
 
+    def _protect_head_size(self, messages: List[Dict[str, Any]]) -> int:
+        """Total count of head messages to protect.
+
+        ``protect_first_n`` is defined as *additional* messages protected
+        beyond the system prompt.  The system prompt (if present at index 0)
+        is always implicitly protected — it's load-bearing context that
+        must never be summarised away.  This keeps semantics stable across
+        call paths where the system prompt may or may not be included in
+        the ``messages`` list (e.g. the gateway ``/compress`` handler
+        strips it before calling compress()).
+
+        Examples:
+          protect_first_n=0 → system prompt only (or nothing if no system msg)
+          protect_first_n=3 → system + first 3 non-system messages
+        """
+        head = 0
+        if messages and messages[0].get("role") == "system":
+            head = 1
+        return head + self.protect_first_n
+
     def _align_boundary_backward(self, messages: List[Dict[str, Any]], idx: int) -> int:
         """Pull a compress-end boundary backward to avoid splitting a
         tool_call / result group.
@@ -1343,7 +1363,7 @@ The user has requested that this compaction PRIORITISE preserving all informatio
         skip the LLM call when the transcript is still entirely inside
         the protected head/tail.
         """
-        compress_start = self._align_boundary_forward(messages, self.protect_first_n)
+        compress_start = self._align_boundary_forward(messages, self._protect_head_size(messages))
         compress_end = self._find_tail_cut_by_tokens(messages, compress_start)
         return compress_start < compress_end
 
@@ -1379,7 +1399,7 @@ The user has requested that this compaction PRIORITISE preserving all informatio
         self._last_aux_model_failure_model = None
         n_messages = len(messages)
         # Only need head + 3 tail messages minimum (token budget decides the real tail size)
-        _min_for_compress = self.protect_first_n + 3 + 1
+        _min_for_compress = self._protect_head_size(messages) + 3 + 1
         if n_messages <= _min_for_compress:
             if not self.quiet_mode:
                 logger.warning(
@@ -1399,7 +1419,7 @@ The user has requested that this compaction PRIORITISE preserving all informatio
             logger.info("Pre-compression: pruned %d old tool result(s)", pruned_count)
 
         # Phase 2: Determine boundaries
-        compress_start = self.protect_first_n
+        compress_start = self._protect_head_size(messages)
         compress_start = self._align_boundary_forward(messages, compress_start)
 
         # Use token-budget tail protection instead of fixed message count