diff --git a/agent/agent_init.py b/agent/agent_init.py
index b839b35576..d8a5bd75e9 100644
--- a/agent/agent_init.py
+++ b/agent/agent_init.py
@@ -520,7 +520,7 @@ def init_agent(
     command: str = None,
     args: list[str] | None = None,
     model: str = "",
-    max_iterations: int = 90,  # Default tool-calling iterations (shared with subagents)
+    max_iterations: int = sys.maxsize,  # Default: unlimited tool-calling iterations (shared with subagents)
     enabled_toolsets: List[str] = None,
     disabled_toolsets: List[str] = None,
     save_trajectories: bool = False,
diff --git a/contributors/emails/fattchris@users.noreply.github.com b/contributors/emails/fattchris@users.noreply.github.com
new file mode 100644
index 0000000000..3c4df87dfe
--- /dev/null
+++ b/contributors/emails/fattchris@users.noreply.github.com
@@ -0,0 +1 @@
+fattchris
diff --git a/hermes_cli/config.py b/hermes_cli/config.py
index 77731bfc3d..5b166427a7 100644
--- a/hermes_cli/config.py
+++ b/hermes_cli/config.py
@@ -3092,29 +3092,35 @@ TURN_LIMIT_UNLIMITED = sys.maxsize
 
 # String spellings that mean "no limit".  Lowercased, whitespace-stripped
 # before comparison so ``"None"``, ``" unlimited "`` etc. all match.
-_UNLIMITED_SPELLINGS = frozenset({"none", "unlimited", "infinite", "∞", "-1", "0"})
+_UNLIMITED_SPELLINGS = frozenset({
+    "none", "null", "unlimited", "infinite", "infinity", "inf",
+    "∞", "-1", "0",
+})
 
 
-def resolve_turn_limit(raw: Any, default: int = 90) -> int:
+def resolve_turn_limit(raw: Any, default: int = TURN_LIMIT_UNLIMITED) -> int:
     """Normalize a raw ``agent.max_turns`` value into an int iteration cap.
 
     Accepts:
-      - ``int`` / ``float`` → ``int(raw)`` (floats truncated; negatives rejected
-        → fall through to ``default``).
+      - ``int`` / ``float`` → ``int(raw)`` (floats truncated; values ≤ 0 mean
+        "no limit" → :data:`TURN_LIMIT_UNLIMITED`).
       - numeric string (``"120"``) → ``int(raw)``.
-      - ``"none"`` / ``"unlimited"`` / ``"infinite"`` / ``"-1"`` / ``"0"``
+      - ``"none"`` / ``"null"`` / ``"unlimited"`` / ``"infinite"`` /
+        ``"infinity"`` / ``"inf"`` / ``"∞"`` / ``"-1"`` / ``"0"``
         (case-insensitive, whitespace-tolerant) → :data:`TURN_LIMIT_UNLIMITED`.
-      - YAML ``None`` / ``null`` (the value is explicitly absent) → ``default``.
+      - YAML ``None`` / ``null`` / absent value → ``default`` (which is itself
+        :data:`TURN_LIMIT_UNLIMITED` — max_turns is unlimited by default).
       - Anything unparseable → ``default`` (with a debug log).
 
     The returned int is always ≥ 1, so loop conditions like
     ``while api_call_count < agent.max_iterations`` behave correctly even when
-    the default path is taken.
+    the default (unlimited) path is taken.
 
     This is the single normalization point for the turn-limit value type.
     Config-reading sites (cli.py, gateway/run.py, cron/scheduler.py) call this
     instead of bare ``int(...)``, so ``agent.max_turns: none`` in config.yaml
-    becomes a first-class supported spelling of "unlimited".
+    becomes a first-class supported spelling of "unlimited". max_turns is
+    unlimited unless the user sets an explicit positive integer cap.
     """
     if raw is None:
         return default
diff --git a/hermes_cli/config_defaults.py b/hermes_cli/config_defaults.py
index 5a3ac68b4f..e20450b98b 100644
--- a/hermes_cli/config_defaults.py
+++ b/hermes_cli/config_defaults.py
@@ -43,7 +43,11 @@ DEFAULT_CONFIG = {
         "terminal_continue": True,
     },
     "agent": {
-        "max_turns": 500,
+        # Unlimited by default. The agent turn cap caused more problems than
+        # it solved (silent mid-task truncation). null = unlimited; set a
+        # positive integer to cap, or use "none"/"unlimited"/"inf"/0/-1 —
+        # all normalized by hermes_cli.config.resolve_turn_limit.
+        "max_turns": None,
         # Optional wall-clock budget in seconds per conversation run.
         # null/absent = feature fully off (zero behavior change). When set,
         # the agent gets a one-time wrap-up notice at 80% elapsed and
diff --git a/run_agent.py b/run_agent.py
index d3588f8f48..f5c4de1274 100644
--- a/run_agent.py
+++ b/run_agent.py
@@ -443,7 +443,7 @@ class AIAgent:
         command: str = None,
         args: list[str] | None = None,
         model: str = "",
-        max_iterations: int = 90,  # Default tool-calling iterations (shared with subagents)
+        max_iterations: int = sys.maxsize,  # Default: unlimited tool-calling iterations (shared with subagents)
         tool_delay: float = None,  # Deprecated: accepted for compatibility, ignored
         enabled_toolsets: List[str] = None,
         disabled_toolsets: List[str] = None,
diff --git a/tests/cli/test_cli_init.py b/tests/cli/test_cli_init.py
index 15e6c653be..cca40f831e 100644
--- a/tests/cli/test_cli_init.py
+++ b/tests/cli/test_cli_init.py
@@ -73,10 +73,14 @@ def _make_cli(env_overrides=None, config_overrides=None, **kwargs):
 class TestMaxTurnsResolution:
     """max_turns must always resolve to a positive integer, never None."""
 
-    def test_default_max_turns_is_integer(self):
+    def test_default_max_turns_is_unlimited(self):
+        # Default is now unlimited (max_turns caused more problems than it
+        # solved). Still a positive int (the sys.maxsize sentinel), so loop
+        # conditions like `count < max_iterations` keep working.
+        import sys
         cli = _make_cli()
         assert isinstance(cli.max_turns, int)
-        assert cli.max_turns == 500
+        assert cli.max_turns == sys.maxsize
 
     def test_explicit_max_turns_honored(self):
         cli = _make_cli(max_turns=25)
diff --git a/tests/hermes_cli/test_resolve_turn_limit.py b/tests/hermes_cli/test_resolve_turn_limit.py
index 417535ffb6..2cacb9d08d 100644
--- a/tests/hermes_cli/test_resolve_turn_limit.py
+++ b/tests/hermes_cli/test_resolve_turn_limit.py
@@ -60,39 +60,40 @@ class TestUnlimitedSpellings:
 
 class TestAbsentAndDefault:
     def test_none_returns_default(self):
-        assert resolve_turn_limit(None) == 90
+        # Default is now unlimited (max_turns caused more problems than it solved).
+        assert resolve_turn_limit(None) == TURN_LIMIT_UNLIMITED
 
     def test_none_custom_default(self):
         assert resolve_turn_limit(None, default=500) == 500
 
     def test_empty_string_returns_default(self):
-        assert resolve_turn_limit("") == 90
+        assert resolve_turn_limit("") == TURN_LIMIT_UNLIMITED
 
     def test_whitespace_only_returns_default(self):
-        assert resolve_turn_limit("   ") == 90
+        assert resolve_turn_limit("   ") == TURN_LIMIT_UNLIMITED
 
     def test_absent_env_var_returns_default(self):
         """Simulates os.getenv() returning None when HERMES_MAX_ITERATIONS unset."""
-        assert resolve_turn_limit(None) == 90
+        assert resolve_turn_limit(None) == TURN_LIMIT_UNLIMITED
 
 
 class TestInvalidInputs:
     def test_bool_rejected(self):
         # bool is an int subclass — must not silently become 1/0
-        assert resolve_turn_limit(True) == 90
-        assert resolve_turn_limit(False) == 90
+        assert resolve_turn_limit(True) == TURN_LIMIT_UNLIMITED
+        assert resolve_turn_limit(False) == TURN_LIMIT_UNLIMITED
 
     def test_garbage_string_returns_default(self):
-        assert resolve_turn_limit("garbage") == 90
-        assert resolve_turn_limit("not_a_number") == 90
+        assert resolve_turn_limit("garbage") == TURN_LIMIT_UNLIMITED
+        assert resolve_turn_limit("not_a_number") == TURN_LIMIT_UNLIMITED
 
     def test_list_returns_default(self):
-        assert resolve_turn_limit([]) == 90
-        assert resolve_turn_limit([90]) == 90
+        assert resolve_turn_limit([]) == TURN_LIMIT_UNLIMITED
+        assert resolve_turn_limit([90]) == TURN_LIMIT_UNLIMITED
 
     def test_dict_returns_default(self):
-        assert resolve_turn_limit({}) == 90
-        assert resolve_turn_limit({"max_turns": 90}) == 90
+        assert resolve_turn_limit({}) == TURN_LIMIT_UNLIMITED
+        assert resolve_turn_limit({"max_turns": 90}) == TURN_LIMIT_UNLIMITED
 
 
 class TestSentinelProperties:
diff --git a/website/docs/user-guide/configuration.md b/website/docs/user-guide/configuration.md
index f64672c850..13430b9ad4 100644
--- a/website/docs/user-guide/configuration.md
+++ b/website/docs/user-guide/configuration.md
@@ -1023,14 +1023,13 @@ Instead, when the budget is actually exhausted (500/500), Hermes injects one mes
 
 ```yaml
 agent:
-  max_turns: 90                # Max iterations per conversation turn (default: 90)
-                               # Set to "none", "unlimited", or 0 for no limit
+  max_turns: none              # Iterations per conversation turn (default: none = unlimited)
+                               # Set a positive integer to cap; "none"/"null"/
+                               # "unlimited"/"inf"/"infinity"/"infinite"/0/-1 = no limit
   api_max_retries: 3           # Retries per provider before fallback engages (default: 3)
 ```
 
-When the iteration budget is fully exhausted, the CLI shows a notification to the user: `⚠ Iteration budget reached (500/500) — response may be incomplete`.
-
-`agent.max_turns` accepts the following "unlimited" spellings (case-insensitive): `"none"`, `"unlimited"`, `"infinite"`, `0`, `-1`. These resolve to a sentinel value (`sys.maxsize`) so the loop never exits on a turn count. Use this for long-running autonomous sessions where the turn budget would otherwise interrupt work.
+`agent.max_turns` is **unlimited by default** — the turn cap caused more problems than it solved (silent mid-task truncation), so out of the box Hermes runs a conversation turn to completion. To impose a cap, set a positive integer. To be explicit about "no limit", any of these case-insensitive spellings work: `"none"`, `"null"`, `"unlimited"`, `"infinite"`, `"infinity"`, `"inf"`, `0`, `-1` (they resolve to a `sys.maxsize` sentinel so the loop never exits on a turn count).
 
 `agent.api_max_retries` controls how many times Hermes retries a provider API call on transient errors (rate limits, connection drops, 5xx) **before** fallback-provider switching engages. The default is `3` — four attempts total. If you have [fallback providers](/user-guide/features/fallback-providers) configured and want to fail over faster, drop this to `0` so the first transient error on your primary immediately hands off to the fallback instead of churning retries against the flaky endpoint.