Expose coordinator retry controls

README.md

@@ -109,6 +109,10 @@ team_harness_agent_models:
   codex: "gpt-5.5"
 team_harness_agent_reasoning_efforts:
   codex: "high"
+# Optional coordinator retry controls. Omit to use team-harness defaults.
+# team_harness_max_retries: 8
+# team_harness_retry_base_delay_s: 2.0
+# team_harness_retry_max_delay_s: 60.0
 team_harness_api_base: "https://openrouter.ai/api/v1"
 team_harness_api_key_env: "OPENROUTER_API_KEY"
 ```
@@ -128,6 +132,9 @@ Rules:
 - `team_harness_api_key_env` must be set during coordinator preflight and again in the worker before `TeamHarness(...)`
 - `team_harness_agent_reasoning_efforts` is optional and only affects workers
   whose team-harness template supports a reasoning-effort flag
+- `team_harness_max_retries`, `team_harness_retry_base_delay_s`, and
+  `team_harness_retry_max_delay_s` are optional coordinator retry controls for
+  transient team-harness API/network errors
 
 Workflow config:
 

docs/http-contract.md

@@ -26,6 +26,9 @@ Run response:
     "team_harness_agents": ["codex"],
     "team_harness_agent_models": {"codex": "gpt-5.5"},
     "team_harness_agent_reasoning_efforts": {"codex": "high"},
+    "team_harness_max_retries": null,
+    "team_harness_retry_base_delay_s": null,
+    "team_harness_retry_max_delay_s": null,
     "team_harness_api_base": "https://openrouter.ai/api/v1",
     "team_harness_api_key_env": "OPENROUTER_API_KEY",
     "team_harness_system_prompt_extension": ""

pyproject.toml

@@ -43,7 +43,7 @@ dependencies = [
     "pydantic>=2.11",
     "pyyaml>=6.0.2",
     "rich>=14.0",
-    "team-harness>=0.2.8",
+    "team-harness>=0.2.9",
     "uvicorn[standard]>=0.32.0",
 ]
 

src/loopy_loop/cli.py

@@ -50,6 +50,10 @@
   - "codex"
 team_harness_agent_models: {{}}
 team_harness_agent_reasoning_efforts: {{}}
+# Optional coordinator retry controls. Omit to use team-harness defaults.
+# team_harness_max_retries: 8
+# team_harness_retry_base_delay_s: 2.0
+# team_harness_retry_max_delay_s: 60.0
 team_harness_api_base: "https://openrouter.ai/api/v1"
 team_harness_api_key_env: "OPENROUTER_API_KEY"
 """

src/loopy_loop/config.py

@@ -10,6 +10,7 @@
 from pydantic import ConfigDict
 from pydantic import Field
 from pydantic import field_validator
+from pydantic import model_validator
 from pydantic import ValidationError
 import yaml
 
@@ -90,6 +91,29 @@ class RootConfig(BaseModel):
             "effort flag will use this value."
         ),
     )
+    team_harness_max_retries: int | None = Field(
+        default=None,
+        ge=0,
+        description=(
+            "Optional coordinator retry budget passed to team-harness. Leave null "
+            "to use the installed team-harness default."
+        ),
+    )
+    team_harness_retry_base_delay_s: float | None = Field(
+        default=None,
+        gt=0,
+        description=(
+            "Optional base delay in seconds for team-harness coordinator retry backoff."
+        ),
+    )
+    team_harness_retry_max_delay_s: float | None = Field(
+        default=None,
+        gt=0,
+        description=(
+            "Optional maximum delay in seconds for team-harness coordinator retry "
+            "backoff."
+        ),
+    )
     team_harness_api_base: str = Field(
         default=DEFAULT_API_BASE,
         description="OpenAI-compatible API base URL passed to team-harness.",
@@ -125,6 +149,20 @@ def validate_non_empty_string_mapping(cls, value: dict[str, str]) -> dict[str, s
     def normalize_api_base_value(cls, value: str) -> str:
         return normalize_api_base(value=value)
 
+    @model_validator(mode="after")
+    def validate_retry_delay_bounds(self) -> "RootConfig":
+        if (
+            self.team_harness_retry_base_delay_s is not None
+            and self.team_harness_retry_max_delay_s is not None
+            and self.team_harness_retry_max_delay_s
+            < self.team_harness_retry_base_delay_s
+        ):
+            raise ValueError(
+                "team_harness_retry_max_delay_s must be greater than or equal to "
+                "team_harness_retry_base_delay_s"
+            )
+        return self
+
 
 class RunAfterSuccesses(BaseModel):
     """Cadence rule that unlocks a workflow after successful runs of another."""

src/loopy_loop/harness_runner.py

@@ -55,12 +55,16 @@ def run_harness_iteration(
         raise
     except TeamHarnessError as exc:
         traceback.print_exc()
+        harness_run_id, harness_output_dir = _failure_harness_paths(
+            detail=exc.detail, harness_output_root=harness_output_root
+        )
         return IterationResult(
             success=False,
             text=None,
             error=str(exc),
             error_detail=exc.detail,
-            harness_run_id="",
+            harness_run_id=harness_run_id,
+            harness_output_dir=harness_output_dir,
         )
     except Exception as exc:
         traceback.print_exc()
@@ -90,6 +94,26 @@ def _build_harness_kwargs(
         "cwd": str(repo_root),
         "console_mode": "silent",
     }
+    retry_kwargs = {
+        "max_retries": config_snapshot.team_harness_max_retries,
+        "retry_base_delay_s": config_snapshot.team_harness_retry_base_delay_s,
+        "retry_max_delay_s": config_snapshot.team_harness_retry_max_delay_s,
+    }
+    configured_retry_kwargs = {
+        key: value for key, value in retry_kwargs.items() if value is not None
+    }
+    if configured_retry_kwargs:
+        if _supports_kwargs(
+            harness_factory=harness_factory, names=configured_retry_kwargs.keys()
+        ):
+            kwargs.update(configured_retry_kwargs)
+        else:
+            raise ConfigError(
+                "Installed team-harness does not support coordinator retry "
+                "controls; upgrade team-harness or remove "
+                "team_harness_max_retries, team_harness_retry_base_delay_s, "
+                "and team_harness_retry_max_delay_s from loopy_loop_config.yaml."
+            )
     agent_override_kwargs = {
         "agent_models": config_snapshot.team_harness_agent_models,
         "agent_reasoning_efforts": config_snapshot.team_harness_agent_reasoning_efforts,
@@ -125,6 +149,20 @@ def _supports_kwargs(
     return all(name in signature.parameters for name in names)
 
 
+def _failure_harness_paths(
+    *, detail: dict[str, object] | None, harness_output_root: Path | None
+) -> tuple[str, str]:
+    if not detail:
+        return "", ""
+    run_id_value = detail.get("run_id")
+    output_dir_value = detail.get("session_output_dir")
+    run_id = run_id_value if isinstance(run_id_value, str) else ""
+    output_dir = output_dir_value if isinstance(output_dir_value, str) else ""
+    if not output_dir and run_id and harness_output_root is not None:
+        output_dir = str(harness_output_root / run_id)
+    return run_id, output_dir
+
+
 def write_iteration_artifacts(
     *, iteration_dir: Path, rendered_prompt: str, iteration_result: IterationResult
 ) -> None:

src/loopy_loop/models.py

@@ -36,6 +36,9 @@ class RootConfigSnapshot(BaseModel):
     team_harness_agents: list[str] = Field(...)
     team_harness_agent_models: dict[str, str] = Field(default_factory=dict)
     team_harness_agent_reasoning_efforts: dict[str, str] = Field(default_factory=dict)
+    team_harness_max_retries: int | None = Field(default=None)
+    team_harness_retry_base_delay_s: float | None = Field(default=None)
+    team_harness_retry_max_delay_s: float | None = Field(default=None)
     team_harness_api_base: str = Field(...)
     team_harness_api_key_env: str = Field(...)
     team_harness_system_prompt_extension: str = Field(...)

src/loopy_loop/templates/atomic_bomberman_lol/loopy_loop_config.yaml

@@ -25,6 +25,10 @@ team_harness_agents:
 team_harness_agent_models:
   codex: "gpt-5.5"
 team_harness_agent_reasoning_efforts: {}
+# Optional coordinator retry controls. Omit to use team-harness defaults.
+# team_harness_max_retries: 8
+# team_harness_retry_base_delay_s: 2.0
+# team_harness_retry_max_delay_s: 60.0
 team_harness_api_base: "https://openrouter.ai/api/v1"
 team_harness_api_key_env: "OPENROUTER_API_KEY"
 team_harness_system_prompt_extension: |

src/loopy_loop/templates/inner_outer_eval/loopy_loop_config.yaml

@@ -13,6 +13,10 @@ team_harness_agent_models:
   gemini: "gemini-3-flash"
 team_harness_agent_reasoning_efforts:
   codex: "high"
+# Optional coordinator retry controls. Omit to use team-harness defaults.
+# team_harness_max_retries: 8
+# team_harness_retry_base_delay_s: 2.0
+# team_harness_retry_max_delay_s: 60.0
 team_harness_api_base: "https://openrouter.ai/api/v1"
 team_harness_api_key_env: "OPENROUTER_API_KEY"
 team_harness_system_prompt_extension: |

src/tests/test_config.py

@@ -22,6 +22,9 @@ def test_load_root_config_and_workflows(repo_builder: Any, monkeypatch: Any) ->
     assert root_config.team_harness_api_base == "https://openrouter.ai/api/v1"
     assert root_config.team_harness_agent_models == {}
     assert root_config.team_harness_agent_reasoning_efforts == {}
+    assert root_config.team_harness_max_retries is None
+    assert root_config.team_harness_retry_base_delay_s is None
+    assert root_config.team_harness_retry_max_delay_s is None
     assert root_config.goal == "Ship a minimal working landing page"
     assert [workflow.id for workflow in workflows] == ["goal_check", "planner"]
     assert workflows[0].priority == 0
@@ -61,6 +64,36 @@ def test_load_root_config_uses_goal_file_and_optional_defaults(
     assert root_config.team_harness_system_prompt_extension == ""
 
 
+def test_load_root_config_accepts_team_harness_retry_controls(
+    repo_builder: Any,
+) -> None:
+    repo_root = repo_builder(
+        root_config={
+            "team_harness_max_retries": 8,
+            "team_harness_retry_base_delay_s": 2.0,
+            "team_harness_retry_max_delay_s": 60.0,
+        }
+    )
+
+    root_config = load_root_config(repo_root=repo_root)
+
+    assert root_config.team_harness_max_retries == 8
+    assert root_config.team_harness_retry_base_delay_s == 2.0
+    assert root_config.team_harness_retry_max_delay_s == 60.0
+
+
+def test_load_root_config_rejects_invalid_retry_delay_bounds(repo_builder: Any) -> None:
+    repo_root = repo_builder(
+        root_config={
+            "team_harness_retry_base_delay_s": 10.0,
+            "team_harness_retry_max_delay_s": 1.0,
+        }
+    )
+
+    with pytest.raises(ConfigError, match="team_harness_retry_max_delay_s"):
+        load_root_config(repo_root=repo_root)
+
+
 def test_load_root_config_rejects_inline_goal(repo_builder: Any) -> None:
     repo_root = repo_builder(root_config={"goal": "Inline goals are no longer used"})
 

src/tests/test_harness_runner.py

@@ -106,6 +106,69 @@ async def run(self, task: str) -> TeamHarnessResult:
     }
 
 
+def test_harness_runner_passes_configured_retry_controls(
+    repo_root: Any, snapshot_factory: Any, monkeypatch: Any
+) -> None:
+    monkeypatch.setenv("OPENROUTER_API_KEY", "secret")
+    captured: dict[str, Any] = {}
+
+    class FakeHarness:
+        def __init__(self, **kwargs: Any) -> None:
+            captured.update(kwargs)
+
+        async def run(self, task: str) -> TeamHarnessResult:
+            return TeamHarnessResult(text="done", agents=[], run_id="run-123")
+
+    run_harness_iteration(
+        repo_root=repo_root,
+        config_snapshot=snapshot_factory(
+            team_harness_max_retries=8,
+            team_harness_retry_base_delay_s=2.0,
+            team_harness_retry_max_delay_s=60.0,
+        ),
+        rendered_prompt="rendered prompt",
+        harness_factory=FakeHarness,
+    )
+
+    assert captured["max_retries"] == 8
+    assert captured["retry_base_delay_s"] == 2.0
+    assert captured["retry_max_delay_s"] == 60.0
+
+
+def test_harness_runner_rejects_retry_controls_for_old_harness(
+    repo_root: Any, snapshot_factory: Any, monkeypatch: Any
+) -> None:
+    monkeypatch.setenv("OPENROUTER_API_KEY", "secret")
+
+    class OldHarness:
+        def __init__(
+            self,
+            *,
+            provider: str,
+            model: str,
+            api_base: str,
+            api_key: str,
+            agents: list[str],
+            agent_models: dict[str, str],
+            agent_reasoning_efforts: dict[str, str],
+            system_prompt: str,
+            cwd: str,
+            console_mode: str,
+        ) -> None:
+            pass
+
+        async def run(self, task: str) -> TeamHarnessResult:
+            return TeamHarnessResult(text="done", agents=[], run_id="run-123")
+
+    with pytest.raises(ConfigError, match="coordinator retry controls"):
+        run_harness_iteration(
+            repo_root=repo_root,
+            config_snapshot=snapshot_factory(team_harness_max_retries=8),
+            rendered_prompt="rendered prompt",
+            harness_factory=OldHarness,
+        )
+
+
 def test_harness_runner_passes_none_api_key_for_codex_provider(
     repo_root: Any, snapshot_factory: Any, monkeypatch: Any
 ) -> None:
@@ -255,6 +318,64 @@ async def run(self, task: str) -> TeamHarnessResult:
     assert result.error_detail == detail
 
 
+def test_harness_runner_preserves_failed_harness_paths_from_error_detail(
+    repo_root: Any, snapshot_factory: Any, monkeypatch: Any, tmp_path: Any
+) -> None:
+    monkeypatch.setenv("OPENROUTER_API_KEY", "secret")
+    detail: dict[str, object] = {
+        "kind": "coordinator_api",
+        "run_id": "run-456",
+        "session_output_dir": str(tmp_path / "outputs" / "0002_inner" / "run-456"),
+    }
+
+    class FakeHarness:
+        def __init__(self, **kwargs: Any) -> None:
+            self.kwargs = kwargs
+
+        async def run(self, task: str) -> TeamHarnessResult:
+            raise TeamHarnessError("boom", detail=detail)
+
+    result = run_harness_iteration(
+        repo_root=repo_root,
+        config_snapshot=snapshot_factory(),
+        rendered_prompt="rendered prompt",
+        harness_output_root=tmp_path / "outputs" / "0002_inner",
+        harness_factory=FakeHarness,
+    )
+
+    assert result.success is False
+    assert result.harness_run_id == "run-456"
+    assert result.harness_output_dir == str(
+        tmp_path / "outputs" / "0002_inner" / "run-456"
+    )
+
+
+def test_harness_runner_derives_failed_output_dir_from_run_id(
+    repo_root: Any, snapshot_factory: Any, monkeypatch: Any, tmp_path: Any
+) -> None:
+    monkeypatch.setenv("OPENROUTER_API_KEY", "secret")
+
+    class FakeHarness:
+        def __init__(self, **kwargs: Any) -> None:
+            self.kwargs = kwargs
+
+        async def run(self, task: str) -> TeamHarnessResult:
+            raise TeamHarnessError("boom", detail={"run_id": "run-789"})
+
+    result = run_harness_iteration(
+        repo_root=repo_root,
+        config_snapshot=snapshot_factory(),
+        rendered_prompt="rendered prompt",
+        harness_output_root=tmp_path / "outputs" / "0003_inner",
+        harness_factory=FakeHarness,
+    )
+
+    assert result.harness_run_id == "run-789"
+    assert result.harness_output_dir == str(
+        tmp_path / "outputs" / "0003_inner" / "run-789"
+    )
+
+
 def test_harness_runner_writes_failure_artifacts(tmp_path: Any) -> None:
     write_iteration_artifacts(
         iteration_dir=tmp_path,