diff --git a/src/bcbench/agent/shared/mcp_gateway.py b/src/bcbench/agent/shared/mcp_gateway.py
index 20d14a853..b5cb0aae2 100644
--- a/src/bcbench/agent/shared/mcp_gateway.py
+++ b/src/bcbench/agent/shared/mcp_gateway.py
@@ -19,7 +19,7 @@
 import time
 from http.client import HTTPConnection
 from http.server import BaseHTTPRequestHandler, ThreadingHTTPServer
-from typing import cast
+from typing import cast, override
 from urllib.parse import urlsplit
 
 from bcbench_core.container import ContainerConfig
@@ -236,6 +236,7 @@ def _build_handler(gateway: BcMcpGateway) -> type[BaseHTTPRequestHandler]:
     class _ProxyHandler(BaseHTTPRequestHandler):
         protocol_version = "HTTP/1.1"
 
+        @override
         def log_message(self, format: str, *args: object) -> None:  # match stdlib signature; silence access log
             pass
 
diff --git a/src/bcbench/commands/evaluate.py b/src/bcbench/commands/evaluate.py
index 3c51e6dd6..93dc3d751 100644
--- a/src/bcbench/commands/evaluate.py
+++ b/src/bcbench/commands/evaluate.py
@@ -1,7 +1,7 @@
 import logging
 import random
 from pathlib import Path
-from typing import Annotated, cast
+from typing import Annotated, cast, override
 
 import typer
 from bcbench_core.filesystem import prepare_run_dir
@@ -346,12 +346,15 @@ class MockEvaluationPipeline(EvaluationPipeline[BaseDatasetEntry]):
     It randomly generates different scenarios to test result handling and serialization.
     """
 
+    @override
     def setup_workspace(self, entry: BaseDatasetEntry, repo_path: Path) -> None:
         logger.info("Mock pipeline: Skipping workspace setup")
 
+    @override
     def setup(self, context: EvaluationContext[BaseDatasetEntry]) -> None:
         logger.info("Mock pipeline: Skipping setup")
 
+    @override
     def run_agent(self, context: EvaluationContext[BaseDatasetEntry], agent_runner: AgentRunner[BaseDatasetEntry]) -> None:
         """Generate random agent metrics and experiment configuration."""
         logger.info("Mock pipeline: Generating random metrics and experiment configuration")
@@ -382,6 +385,7 @@ def run_agent(self, context: EvaluationContext[BaseDatasetEntry], agent_runner:
         logger.info(f"Using agent metrics: {context.metrics}")
         logger.info(f"Using experiment configuration: {context.experiment}")
 
+    @override
     def evaluate(self, context: EvaluationContext[BaseDatasetEntry]) -> None:
         """Create random evaluation result to test different outcome scenarios."""
         logger.info("Mock pipeline: Generating random evaluation result")
diff --git a/src/bcbench/dataset/codereview.py b/src/bcbench/dataset/codereview.py
index 55a6242f4..c08b77ebf 100644
--- a/src/bcbench/dataset/codereview.py
+++ b/src/bcbench/dataset/codereview.py
@@ -1,7 +1,7 @@
 from __future__ import annotations
 
 from enum import StrEnum
-from typing import Annotated, Self
+from typing import Annotated, Self, override
 
 from pydantic import BaseModel, ConfigDict, Field, field_validator, model_validator
 
@@ -81,6 +81,7 @@ def _coerce_severity(cls, value: object) -> Severity | None:
     def severity_label(self) -> str:
         return self.severity.value if self.severity is not None else "unspecified"
 
+    @override
     def __str__(self) -> str:
         loc = f"{self.file}:{self.line_start}"
         if self.line_end and self.line_end != self.line_start:
@@ -129,9 +130,11 @@ def _validate_article_annotations(self) -> Self:
             )
         return self
 
+    @override
     def get_task(self) -> str:
         return self.patch
 
+    @override
     def get_expected_output(self) -> str:
         return "\n".join(str(c) for c in self.expected_comments)
 
diff --git a/src/bcbench/dataset/dataset_entry.py b/src/bcbench/dataset/dataset_entry.py
index 0162f7c9f..1310d3d43 100644
--- a/src/bcbench/dataset/dataset_entry.py
+++ b/src/bcbench/dataset/dataset_entry.py
@@ -4,7 +4,7 @@
 import re
 from abc import abstractmethod
 from pathlib import Path
-from typing import Annotated, Literal, Self
+from typing import Annotated, Literal, Self, override
 
 from bcbench_core.dataset import TestEntry
 from pydantic import BaseModel, ConfigDict, Field, model_validator
@@ -118,6 +118,7 @@ class RepoGroundedEntry(BaseDatasetEntry):
     patch: Annotated[str, Field(min_length=1, pattern=r"^[^\x00]*$")]
 
     @property
+    @override
     def customization_profile(self) -> str:
         return self.repo.replace("/", "-")
 
@@ -125,6 +126,7 @@ def customization_profile(self) -> str:
     def problem_statement_dir(self) -> Path:
         return _config.paths.problem_statement_dir / self.instance_id
 
+    @override
     def get_task(self) -> str:
         readme_path = self.problem_statement_dir / _config.file_patterns.problem_statement_readme
         return readme_path.read_text(encoding="utf-8")
@@ -158,6 +160,7 @@ def validate_baseapp_patches_are_w1_only(self) -> Self:
 class BugFixEntry(_BugFixTestGenBase):
     """Dataset entry for the bug-fix category."""
 
+    @override
     def get_expected_output(self) -> str:
         return self.patch
 
@@ -165,6 +168,7 @@ def get_expected_output(self) -> str:
 class TestGenEntry(_BugFixTestGenBase):
     """Dataset entry for the test-generation category."""
 
+    @override
     def get_expected_output(self) -> str:
         return self.test_patch
 
@@ -178,12 +182,15 @@ class NL2ALEntry(BaseDatasetEntry):
     audience: Literal["Business", "Technical", "Both"]
 
     @property
+    @override
     def customization_profile(self) -> str:
         return "nl2al"
 
+    @override
     def get_task(self) -> str:
         return self.nl_prompt
 
+    @override
     def get_expected_output(self) -> Checklist:
         return {"assertions": self.expected}
 
@@ -204,11 +211,14 @@ class DataQueryEntry(BaseDatasetEntry):
     ordered: bool = False
 
     @property
+    @override
     def customization_profile(self) -> str:
         return "dataquery"
 
+    @override
     def get_task(self) -> str:
         return self.nl_prompt
 
+    @override
     def get_expected_output(self) -> str:
         return self.gold_query
diff --git a/src/bcbench/dataset/extensibility_request.py b/src/bcbench/dataset/extensibility_request.py
index aa5a6e58d..3811adee5 100644
--- a/src/bcbench/dataset/extensibility_request.py
+++ b/src/bcbench/dataset/extensibility_request.py
@@ -1,6 +1,6 @@
 from __future__ import annotations
 
-from typing import Annotated, Literal
+from typing import Annotated, Literal, override
 
 from pydantic import Field
 
@@ -20,12 +20,14 @@ class ExtRequestAdvisorEntry(RepoGroundedEntry):
     comments: str = ""
     expected: Annotated[list[ChecklistAssertion], Field(min_length=1)]
 
+    @override
     def get_task(self) -> str:
         sections = [f"# {self.title}", "", self.description.rstrip()]
         if self.comments.strip():
             sections += ["", "## Additional requester context", "", self.comments.rstrip()]
         return "\n".join(sections)
 
+    @override
     def get_expected_output(self) -> Checklist:
         return {"assertions": self.expected}
 
@@ -43,6 +45,7 @@ class ExtRequestImplementEntry(RepoGroundedEntry):
     # LLM-judge checklist: expected event/signature/placement and expected layer propagation.
     expected: Annotated[list[ChecklistAssertion], Field(min_length=1)]
 
+    @override
     def get_expected_output(self) -> Checklist:
         return {"assertions": self.expected}
 
@@ -83,6 +86,7 @@ class ExtRequestTriageEntry(RepoGroundedEntry):
     # LLM-judge checklist: expected labels_to_set, issue_state and advisory-comment substance.
     expected: Annotated[list[ChecklistAssertion], Field(min_length=1)]
 
+    @override
     def get_task(self) -> str:
         sections = [f"# {self.title}", "", self.description.rstrip()]
         if self.current_labels:
@@ -91,5 +95,6 @@ def get_task(self) -> str:
             sections += ["", "## Follow-up conversation", "", self.comments.rstrip()]
         return "\n".join(sections)
 
+    @override
     def get_expected_output(self) -> Checklist:
         return {"assertions": self.expected}
diff --git a/src/bcbench/evaluate/bugfix.py b/src/bcbench/evaluate/bugfix.py
index c9a403f85..4f513f40f 100644
--- a/src/bcbench/evaluate/bugfix.py
+++ b/src/bcbench/evaluate/bugfix.py
@@ -1,5 +1,6 @@
 import logging
 from pathlib import Path
+from typing import override
 
 from bcbench_core.bc import build_and_publish_projects
 from bcbench_core.exceptions import BuildError, TestExecutionError
@@ -21,11 +22,13 @@
 class BugFixPipeline(EvaluationPipeline[BugFixEntry]):
     """Pipeline for bug-fix evaluation category."""
 
+    @override
     def setup_workspace(self, entry: BugFixEntry, repo_path: Path) -> None:
         setup_repo_prebuild(entry, repo_path)
         copy_problem_statement_folder(entry, repo_path)
         set_runtime_version(repo_path, entry.project_paths)
 
+    @override
     def setup(self, context: EvaluationContext[BugFixEntry]) -> None:
         setup_repo_prebuild(context.entry, context.repo_path)
 
@@ -39,10 +42,12 @@ def setup(self, context: EvaluationContext[BugFixEntry]) -> None:
         copy_problem_statement_folder(context.entry, context.repo_path)
         set_runtime_version(context.repo_path, context.entry.project_paths)
 
+    @override
     def run_agent(self, context: EvaluationContext[BugFixEntry], agent_runner: AgentRunner[BugFixEntry]) -> None:
         with github_log_group(f"{context.agent_name} -- Entry: {context.entry.instance_id}"):
             context.metrics, context.experiment = agent_runner(context)
 
+    @override
     def evaluate(self, context: EvaluationContext[BugFixEntry]) -> None:
         container = context.get_container()
         test_projects, _app_projects = categorize_projects(context.entry.project_paths)
diff --git a/src/bcbench/evaluate/codereview.py b/src/bcbench/evaluate/codereview.py
index bbc72797f..8ad821415 100644
--- a/src/bcbench/evaluate/codereview.py
+++ b/src/bcbench/evaluate/codereview.py
@@ -1,6 +1,7 @@
 import logging
 import subprocess
 from pathlib import Path
+from typing import override
 
 from bcbench_core.git import apply_patch, fetch_commit_if_missing
 
@@ -31,6 +32,7 @@ class CodeReviewPipeline(EvaluationPipeline[CodeReviewEntry]):
     as local git changes so the agent can review the branch diff directly.
     """
 
+    @override
     def setup_workspace(self, entry: CodeReviewEntry, repo_path: Path) -> None:
         """Setup workspace for code review by applying the entry patch as local changes."""
         # Code-review base commits are pre-squash PR commits, so they might be missing from local dev setups.
@@ -48,13 +50,16 @@ def setup_workspace(self, entry: CodeReviewEntry, repo_path: Path) -> None:
                 check=True,
             )
 
+    @override
     def setup(self, context: EvaluationContext[CodeReviewEntry]) -> None:
         self.setup_workspace(context.entry, context.repo_path)
 
+    @override
     def run_agent(self, context: EvaluationContext[CodeReviewEntry], agent_runner: AgentRunner[CodeReviewEntry]) -> None:
         with github_log_group(f"{context.agent_name} -- Entry: {context.entry.instance_id}"):
             context.metrics, context.experiment = agent_runner(context)
 
+    @override
     def evaluate(self, context: EvaluationContext[CodeReviewEntry]) -> None:
         review_output_file: Path = context.repo_path / REVIEW_OUTPUT_FILE
 
diff --git a/src/bcbench/evaluate/dataquery.py b/src/bcbench/evaluate/dataquery.py
index e5ded1c93..61df1a724 100644
--- a/src/bcbench/evaluate/dataquery.py
+++ b/src/bcbench/evaluate/dataquery.py
@@ -3,6 +3,7 @@
 from collections.abc import Callable, Mapping, Sequence
 from decimal import Decimal, InvalidOperation
 from pathlib import Path
+from typing import override
 
 from bcbench_core.filesystem import clear_directory
 
@@ -94,17 +95,21 @@ class DataQueryPipeline(EvaluationPipeline[DataQueryEntry]):
     genuinely querying the environment.
     """
 
+    @override
     def setup_workspace(self, entry: DataQueryEntry, repo_path: Path) -> None:
         # The workspace is shared into the running container, so its contents are cleared in place.
         clear_directory(repo_path)
 
+    @override
     def setup(self, context: EvaluationContext[DataQueryEntry]) -> None:
         self.setup_workspace(context.entry, context.repo_path)
 
+    @override
     def run_agent(self, context: EvaluationContext[DataQueryEntry], agent_runner: Callable) -> None:
         with github_log_group(f"{context.agent_name} -- Entry: {context.entry.instance_id}"):
             context.metrics, context.experiment = agent_runner(context)
 
+    @override
     def evaluate(self, context: EvaluationContext[DataQueryEntry]) -> None:
         query_file = context.repo_path / GENERATED_QUERY_FILE
         # query.al is only an inspection artifact now; scoring is on the data the agent retrieved.
diff --git a/src/bcbench/evaluate/ext_request_advisor.py b/src/bcbench/evaluate/ext_request_advisor.py
index 9fc798c5a..7c98a12d0 100644
--- a/src/bcbench/evaluate/ext_request_advisor.py
+++ b/src/bcbench/evaluate/ext_request_advisor.py
@@ -1,5 +1,6 @@
 import logging
 from pathlib import Path
+from typing import override
 
 from bcbench.dataset import ExtRequestAdvisorEntry
 from bcbench.evaluate.base import AgentRunner, EvaluationPipeline
@@ -18,17 +19,21 @@
 class ExtRequestAdvisorPipeline(EvaluationPipeline[ExtRequestAdvisorEntry]):
     """Offline single-shot proxy for the interactive extensibility advisor."""
 
+    @override
     def setup_workspace(self, entry: ExtRequestAdvisorEntry, repo_path: Path) -> None:
         setup_repo_prebuild(entry, repo_path)
         (repo_path / ADVISOR_RESULT_FILE).unlink(missing_ok=True)
 
+    @override
     def setup(self, context: EvaluationContext[ExtRequestAdvisorEntry]) -> None:
         self.setup_workspace(context.entry, context.repo_path)
 
+    @override
     def run_agent(self, context: EvaluationContext[ExtRequestAdvisorEntry], agent_runner: AgentRunner[ExtRequestAdvisorEntry]) -> None:
         with github_log_group(f"{context.agent_name} -- Entry: {context.entry.instance_id}"):
             context.metrics, context.experiment = agent_runner(context)
 
+    @override
     def evaluate(self, context: EvaluationContext[ExtRequestAdvisorEntry]) -> None:
         result_path = context.repo_path / ADVISOR_RESULT_FILE
         raw = result_path.read_text(encoding="utf-8").strip() if result_path.exists() else ""
diff --git a/src/bcbench/evaluate/ext_request_implement.py b/src/bcbench/evaluate/ext_request_implement.py
index e95c91541..79d6ee20c 100644
--- a/src/bcbench/evaluate/ext_request_implement.py
+++ b/src/bcbench/evaluate/ext_request_implement.py
@@ -1,5 +1,6 @@
 import logging
 from pathlib import Path
+from typing import override
 
 from bcbench_core.exceptions import EmptyDiffError
 from bcbench_core.git import stage_and_get_diff
@@ -24,18 +25,22 @@ class ExtRequestImplementPipeline(EvaluationPipeline[ExtRequestImplementEntry]):
     by an LLM judge against the entry checklist.
     """
 
+    @override
     def setup_workspace(self, entry: ExtRequestImplementEntry, repo_path: Path) -> None:
         setup_repo_prebuild(entry, repo_path)
         copy_problem_statement_folder(entry, repo_path)
         set_runtime_version(repo_path, entry.project_paths)
 
+    @override
     def setup(self, context: EvaluationContext[ExtRequestImplementEntry]) -> None:
         self.setup_workspace(context.entry, context.repo_path)
 
+    @override
     def run_agent(self, context: EvaluationContext[ExtRequestImplementEntry], agent_runner: AgentRunner[ExtRequestImplementEntry]) -> None:
         with github_log_group(f"{context.agent_name} -- Entry: {context.entry.instance_id}"):
             context.metrics, context.experiment = agent_runner(context)
 
+    @override
     def evaluate(self, context: EvaluationContext[ExtRequestImplementEntry]) -> None:
         try:
             generated_patch = stage_and_get_diff(context.repo_path, exclude=("**/app.json", "*.docx", "*.md"))
diff --git a/src/bcbench/evaluate/ext_request_triage.py b/src/bcbench/evaluate/ext_request_triage.py
index d0dfdc562..1743916e8 100644
--- a/src/bcbench/evaluate/ext_request_triage.py
+++ b/src/bcbench/evaluate/ext_request_triage.py
@@ -10,6 +10,7 @@
 
 import logging
 from pathlib import Path
+from typing import override
 
 from bcbench.dataset import ExtRequestTriageEntry
 from bcbench.evaluate.base import AgentRunner, EvaluationPipeline
@@ -28,17 +29,21 @@
 class ExtRequestTriagePipeline(EvaluationPipeline[ExtRequestTriageEntry]):
     """Pipeline for the extensibility-request-triage category — no BC container, no build, no tests."""
 
+    @override
     def setup_workspace(self, entry: ExtRequestTriageEntry, repo_path: Path) -> None:
         setup_repo_prebuild(entry, repo_path)
         (repo_path / TRIAGE_RESULT_FILE).unlink(missing_ok=True)
 
+    @override
     def setup(self, context: EvaluationContext[ExtRequestTriageEntry]) -> None:
         self.setup_workspace(context.entry, context.repo_path)
 
+    @override
     def run_agent(self, context: EvaluationContext[ExtRequestTriageEntry], agent_runner: AgentRunner[ExtRequestTriageEntry]) -> None:
         with github_log_group(f"{context.agent_name} -- Entry: {context.entry.instance_id}"):
             context.metrics, context.experiment = agent_runner(context)
 
+    @override
     def evaluate(self, context: EvaluationContext[ExtRequestTriageEntry]) -> None:
         result_path = context.repo_path / TRIAGE_RESULT_FILE
         raw = result_path.read_text(encoding="utf-8").strip() if result_path.exists() else ""
diff --git a/src/bcbench/evaluate/nl2al.py b/src/bcbench/evaluate/nl2al.py
index 2848f6fe3..e04e4c65b 100644
--- a/src/bcbench/evaluate/nl2al.py
+++ b/src/bcbench/evaluate/nl2al.py
@@ -2,6 +2,7 @@
 import os
 import subprocess
 from pathlib import Path
+from typing import override
 
 from bcbench_core.artifacts import copy_symbol_apps
 from bcbench_core.exceptions import EmptyDiffError
@@ -56,20 +57,24 @@ def _git_init_and_commit(repo_path: Path) -> None:
 class NL2ALPipeline(EvaluationPipeline[NL2ALEntry]):
     """Pipeline for NL2AL evaluation category — generate AL code from natural language."""
 
+    @override
     def setup_workspace(self, entry: NL2ALEntry, repo_path: Path) -> None:
         _reset_repo_path(repo_path)
         copy_symbol_apps(repo_path / entry.project_paths[0], entry.environment_setup_version)
         _git_init_and_commit(repo_path)
 
+    @override
     def setup(self, context: EvaluationContext[NL2ALEntry]) -> None:
         self.setup_workspace(context.entry, context.repo_path)
 
+    @override
     def run_agent(self, context: EvaluationContext[NL2ALEntry], agent_runner: AgentRunner[NL2ALEntry]) -> None:
         # Single attempt — retries are disabled. An empty diff (the agent asked for clarification
         # instead of editing) is scored as a failure in evaluate(), not re-run.
         with github_log_group(f"{context.agent_name} -- Entry: {context.entry.instance_id}"):
             context.metrics, context.experiment = agent_runner(context)
 
+    @override
     def evaluate(self, context: EvaluationContext[NL2ALEntry]) -> None:
         try:
             generated_patch = stage_and_get_diff(context.repo_path, exclude=("**/app.json", "*.docx", "*.md"))
diff --git a/src/bcbench/evaluate/testgeneration.py b/src/bcbench/evaluate/testgeneration.py
index 3ccef67eb..2e8c3f45e 100644
--- a/src/bcbench/evaluate/testgeneration.py
+++ b/src/bcbench/evaluate/testgeneration.py
@@ -1,5 +1,6 @@
 import logging
 from pathlib import Path
+from typing import override
 
 import yaml
 from bcbench_core.bc import build_and_publish_projects, run_test_suite
@@ -54,11 +55,13 @@ def _apply_input_postbuild(self, entry: TestGenEntry, repo_path: Path) -> None:
             case _:
                 raise ValueError(f"Unhandled test generation input mode: {input_mode}")
 
+    @override
     def setup_workspace(self, entry: TestGenEntry, repo_path: Path) -> None:
         setup_repo_prebuild(entry, repo_path)
         self._apply_input_postbuild(entry, repo_path)
         set_runtime_version(repo_path, entry.project_paths)
 
+    @override
     def setup(self, context: EvaluationContext[TestGenEntry]) -> None:
         setup_repo_prebuild(context.entry, context.repo_path)
 
@@ -72,10 +75,12 @@ def setup(self, context: EvaluationContext[TestGenEntry]) -> None:
         self._apply_input_postbuild(context.entry, context.repo_path)
         set_runtime_version(context.repo_path, context.entry.project_paths)
 
+    @override
     def run_agent(self, context: EvaluationContext[TestGenEntry], agent_runner: AgentRunner[TestGenEntry]) -> None:
         with github_log_group(f"{context.agent_name} -- Entry: {context.entry.instance_id}"):
             context.metrics, context.experiment = agent_runner(context)
 
+    @override
     def evaluate(self, context: EvaluationContext[TestGenEntry]) -> None:
         container = context.get_container()
         test_projects, app_projects = categorize_projects(context.entry.project_paths)
diff --git a/src/bcbench/results/base.py b/src/bcbench/results/base.py
index 4d113dcd3..165acb0fe 100644
--- a/src/bcbench/results/base.py
+++ b/src/bcbench/results/base.py
@@ -3,7 +3,7 @@
 import json
 import logging
 from pathlib import Path
-from typing import Any, Self, cast
+from typing import Any, Self, cast, override
 
 from pydantic import BaseModel, model_validator
 
@@ -126,12 +126,14 @@ def create_result(cls, context: "EvaluationContext", output: str, *, build: bool
         return cls(**cls._base_fields(context), output=output, build=build, resolved=resolved, error_message=error_message)
 
     @property
+    @override
     def status_label(self) -> str:
         if self.timeout:
             return "Timeout"
         return "Success" if self.resolved else "Failed"
 
     @property
+    @override
     def category_metrics(self) -> dict[str, int | float | bool]:
         return {"resolved": self.resolved, "build": self.build}
 
@@ -158,10 +160,12 @@ def restore_missing_timeout_judge_model(cls, payload: object) -> object:
         return {**payload, "judge_model": judge_model}
 
     @classmethod
+    @override
     def _base_fields(cls, context: "EvaluationContext") -> dict[str, Any]:
         return {**super()._base_fields(context), "judge_model": context.category.judge_model}
 
     @property
+    @override
     def export_metadata(self) -> dict[str, str | int | float | bool | None]:
         return {"judge_model": self.judge_model}
 
@@ -186,6 +190,7 @@ def create_empty_output(cls, context: "EvaluationContext") -> Self:
         return cls(**cls._base_fields(context), output="")
 
     @property
+    @override
     def status_label(self) -> str:
         if self.timeout:
             return "Timeout"
diff --git a/src/bcbench/results/codereview.py b/src/bcbench/results/codereview.py
index 89f8bb60a..80ae8f3e9 100644
--- a/src/bcbench/results/codereview.py
+++ b/src/bcbench/results/codereview.py
@@ -1,5 +1,5 @@
 from collections.abc import Sequence
-from typing import NamedTuple, Self
+from typing import NamedTuple, Self, override
 
 import numpy as np
 from bcbench_core.scoring import f1_score, f_beta_score, precision_recall
@@ -233,6 +233,7 @@ def create_invalid(
         )
 
     @property
+    @override
     def category_metrics(self) -> dict[str, int | float | bool]:
         return {
             "generated_comment_count": len(self.generated_comments),
@@ -251,6 +252,7 @@ def category_metrics(self) -> dict[str, int | float | bool]:
         }
 
     @property
+    @override
     def display_row(self) -> dict[str, str]:
         return {
             "Generated": str(len(self.generated_comments)),
@@ -315,6 +317,7 @@ def metric(value: float | None, digits: int = 1) -> str:
             "\n"
         )
 
+    @override
     def render_github_metrics_markdown(self) -> str:
         micro_p = self.precision * 100
         micro_r = self.recall * 100
@@ -356,6 +359,7 @@ def render_github_metrics_markdown(self) -> str:
             f"{_METRIC_EXPLANATIONS}"
         )
 
+    @override
     def render_console_metrics(self) -> RenderableType:
         metric_columns = ["Precision", "Recall", "F1", "Fβ (β=0.5)", "Fβ (β=2)"]
 
@@ -420,6 +424,7 @@ def render_console_metrics(self) -> RenderableType:
         )
 
     @classmethod
+    @override
     def from_results(cls, results: Sequence[BaseEvaluationResult], run_id: str) -> "CodeReviewResultSummary":
         summary = super().from_results(results, run_id)
         assert isinstance(summary, CodeReviewResultSummary)
diff --git a/src/bcbench/results/leaderboard.py b/src/bcbench/results/leaderboard.py
index 782c77e57..b327a59bc 100644
--- a/src/bcbench/results/leaderboard.py
+++ b/src/bcbench/results/leaderboard.py
@@ -4,7 +4,7 @@
 from collections import defaultdict
 from collections.abc import Sequence
 from pathlib import Path
-from typing import Any
+from typing import Any, override
 
 from bcbench_core.scoring import pass_hat_k
 from bcbench_core.stats import bootstrap_ci
@@ -91,6 +91,7 @@ class ExecutionBasedLeaderboardAggregate(LeaderboardAggregate):
     pass_hat_5: float | None = None
 
     @classmethod
+    @override
     def from_runs(cls, runs: Sequence[EvaluationResultSummary]) -> "ExecutionBasedLeaderboardAggregate":
         base = super().from_runs(runs)
         assert isinstance(base, ExecutionBasedLeaderboardAggregate)
@@ -123,6 +124,7 @@ class JudgeBasedLeaderboardAggregate(LeaderboardAggregate):
     judge_model: str
 
     @classmethod
+    @override
     def _base_fields(cls, runs: Sequence[EvaluationResultSummary]) -> dict[str, Any]:
         from bcbench.results.summary import JudgeBasedEvaluationResultSummary
 
@@ -157,6 +159,7 @@ class CodeReviewLeaderboardAggregate(JudgeBasedLeaderboardAggregate):
     average_ai_credits: float | None = None
 
     @classmethod
+    @override
     def from_runs(cls, runs: Sequence[EvaluationResultSummary]) -> "CodeReviewLeaderboardAggregate":
         from bcbench.results.codereview import CodeReviewResultSummary
 
diff --git a/src/bcbench/results/summary.py b/src/bcbench/results/summary.py
index eea4d3e6f..ecff8e365 100644
--- a/src/bcbench/results/summary.py
+++ b/src/bcbench/results/summary.py
@@ -7,7 +7,7 @@
 from datetime import UTC, date, datetime
 from importlib.metadata import PackageNotFoundError, version
 from pathlib import Path
-from typing import TYPE_CHECKING, Any
+from typing import TYPE_CHECKING, Any, override
 
 from pydantic import BaseModel, Field
 
@@ -164,10 +164,12 @@ class ExecutionBasedEvaluationResultSummary(EvaluationResultSummary):
     # Per-instance pass/fail for aggregate metrics (pass^k, CI)
     instance_results: dict[str, bool] = Field(default_factory=dict)
 
+    @override
     def render_github_metrics_markdown(self) -> str:
         return f"## Result Summary\n- Resolved: {self.resolved}\n- Failed: {self.failed}\n- Build: {self.build}\n- Pass Rate: {self.percentage}%\n"
 
     @classmethod
+    @override
     def from_results(cls, results: Sequence[BaseEvaluationResult], run_id: str) -> "ExecutionBasedEvaluationResultSummary":
         from bcbench.results.base import ExecutionBasedEvaluationResult
 
@@ -196,6 +198,7 @@ class JudgeBasedEvaluationResultSummary(EvaluationResultSummary):
     judge_model: str
 
     @classmethod
+    @override
     def _base_fields(cls, results: Sequence[BaseEvaluationResult], run_id: str) -> dict[str, Any]:
         from bcbench.results.base import JudgeScoredEvaluationResult
 
@@ -203,10 +206,12 @@ def _base_fields(cls, results: Sequence[BaseEvaluationResult], run_id: str) -> d
         assert isinstance(first_result, JudgeScoredEvaluationResult)
         return {**super()._base_fields(results, run_id), "judge_model": first_result.judge_model}
 
+    @override
     def combination_key(self) -> tuple[str | None, ...]:
         """Runs judged by different models are aggregated separately."""
         return (*super().combination_key(), self.judge_model)
 
+    @override
     def render_github_metrics_markdown(self) -> str:
         """Judge scoring happens externally, so there are no in-run metrics to surface."""
         return ""
diff --git a/src/bcbench/results/testgeneration.py b/src/bcbench/results/testgeneration.py
index a703752ed..24b5153df 100644
--- a/src/bcbench/results/testgeneration.py
+++ b/src/bcbench/results/testgeneration.py
@@ -1,4 +1,4 @@
-from typing import Self
+from typing import Self, override
 
 from bcbench.results.base import ExecutionBasedEvaluationResult
 from bcbench.types import EvaluationContext
@@ -11,10 +11,12 @@ class TestGenerationResult(ExecutionBasedEvaluationResult):
     post_patch_passed: bool = False
 
     @property
+    @override
     def category_metrics(self) -> dict[str, int | float | bool]:
         return {**super().category_metrics, "pre_patch_failed": self.pre_patch_failed, "post_patch_passed": self.post_patch_passed}
 
     @property
+    @override
     def display_row(self) -> dict[str, str]:
         return {
             "Pre-Patch Failed": "Yes" if self.pre_patch_failed else "No",
@@ -22,6 +24,7 @@ def display_row(self) -> dict[str, str]:
         }
 
     @classmethod
+    @override
     def create_success(cls, context: "EvaluationContext", output: str) -> Self:
         return cls(**cls._base_fields(context), output=output, resolved=True, build=True, pre_patch_failed=True, post_patch_passed=True)
 
diff --git a/tests/test_evaluate_pipeline.py b/tests/test_evaluate_pipeline.py
index 131437b04..00cd01890 100644
--- a/tests/test_evaluate_pipeline.py
+++ b/tests/test_evaluate_pipeline.py
@@ -2,6 +2,7 @@
 
 import json
 from pathlib import Path
+from typing import override
 from unittest.mock import patch
 
 import pytest
@@ -24,18 +25,22 @@ def __init__(self, *, raise_in_evaluate: Exception | None = None, raise_in_run_a
         self.run_agent_called = False
         self.evaluate_called = False
 
+    @override
     def setup_workspace(self, entry: E, repo_path: Path) -> None:
         pass
 
+    @override
     def setup(self, context: EvaluationContext[E]) -> None:
         self.setup_called = True
 
+    @override
     def run_agent(self, context: EvaluationContext[E], agent_runner: AgentRunner[E]) -> None:
         self.run_agent_called = True
         if self.raise_in_run_agent is not None:
             raise self.raise_in_run_agent
         context.metrics, context.experiment = agent_runner(context)
 
+    @override
     def evaluate(self, context: EvaluationContext[E]) -> None:
         self.evaluate_called = True
         if self.raise_in_evaluate is not None:
diff --git a/tests/test_mcp_gateway.py b/tests/test_mcp_gateway.py
index 1d00234b1..d2e718f25 100644
--- a/tests/test_mcp_gateway.py
+++ b/tests/test_mcp_gateway.py
@@ -4,6 +4,7 @@
 import time
 from http.client import HTTPConnection
 from http.server import BaseHTTPRequestHandler, ThreadingHTTPServer
+from typing import override
 from urllib.parse import urlsplit
 
 import pytest
@@ -33,6 +34,7 @@ class _RecordingServer(ThreadingHTTPServer):
 class _UpstreamHandler(BaseHTTPRequestHandler):
     protocol_version = "HTTP/1.1"
 
+    @override
     def log_message(self, format: str, *args: object) -> None:  # match stdlib signature; silence access log
         pass
 
@@ -163,6 +165,7 @@ def test_counts_forwarded_requests(self, gateway):
 class _McpHandler(BaseHTTPRequestHandler):
     protocol_version = "HTTP/1.1"
 
+    @override
     def log_message(self, format: str, *args: object) -> None:  # match stdlib signature; silence access log
         pass
 
@@ -252,6 +255,7 @@ class _EmptyThenToolsHandler(BaseHTTPRequestHandler):
     protocol_version = "HTTP/1.1"
     tools_list_calls = 0
 
+    @override
     def log_message(self, format: str, *args: object) -> None:  # match stdlib signature; silence access log
         pass
 
@@ -307,6 +311,7 @@ class _HeldOpenSseHandler(BaseHTTPRequestHandler):
 
     protocol_version = "HTTP/1.1"
 
+    @override
     def log_message(self, format: str, *args: object) -> None:  # match stdlib signature; silence access log
         pass
 
@@ -325,6 +330,7 @@ class _HeldOpenPostSseHandler(BaseHTTPRequestHandler):
 
     protocol_version = "HTTP/1.1"
 
+    @override
     def log_message(self, format: str, *args: object) -> None:  # match stdlib signature; silence access log
         pass
 
@@ -378,6 +384,7 @@ class _InitializeExperimentalHandler(BaseHTTPRequestHandler):
 
     protocol_version = "HTTP/1.1"
 
+    @override
     def log_message(self, format: str, *args: object) -> None:  # match stdlib signature; silence access log
         pass