headroomlabs-ai/headroom

39.5

Adjusted Score

39.5

Raw Score

100%

Time Factor

2026-07-14

Last Push

59.1K

Stars

Python

Language

551.1K

Lines of Code

1.7K

Files

15.1K

Pattern Hits

2026-07-14

Scan Date

0.11

HC Hit Rate

What These Metrics Mean

Adjusted Score: Primary synthetic code indicator. Raw score normalised per 1,000 lines of code and multiplied by the temporal discount factor. This is the definitive comparative metric — use it to rank repositories by AI authorship density.
Raw Score: The unmodified sum of all severity-weighted, context-multiplied pattern match scores before temporal discounting. Reflects the absolute signal strength independent of when the repository was last active.
Time Factor: The temporal discount multiplier (0–100%) applied to the raw score. Repositories last updated before ChatGPT's launch (Nov 2022) receive a 5% factor. Full signal is only assigned to repositories active in the post-adoption era (Jan 2024+).
Pattern Hits: Total count of individual pattern matches across all files and categories. A high hit count with a low score may indicate a very large codebase with isolated AI snippets; a low count with a high score indicates dense, concentrated AI signatures.
HC Hit Rate: High+Critical pattern hits per file, averaged across the repository. This orthogonal signal catches repositories where a few files are densely packed with high-severity AI tells — a strong indicator even when the normalised score appears moderate due to codebase size.
Lines of Code / Files: Total lines and files analysed. The scanner examines 94 file extensions. These denominators are used to normalise the score, enabling fair comparison between repositories of vastly different sizes.

Score History

This chart maps the temporal evolution of the adjusted synthetic code score across successive scan runs. An upward trajectory indicates ongoing incorporation of AI-generated code or expanding LLM-assisted scaffolding; a stable or declining trajectory may reflect active human refactoring, code removal, or the adoption of stricter authorship policies. The dashed secondary line (right axis) independently tracks total raw pattern hit count, which can diverge from the normalised score when codebase size changes significantly between scans.

Severity Breakdown

Classifies detected patterns by their diagnostic confidence and structural impact. CRITICAL patterns (coefficient 10) represent definitive synthetic signatures — hallucinated imports, explicit LLM attribution metadata — virtually never produced by human authors. HIGH (5) indicates strong structural tells such as cross-file repetition or cross-linguistic idioms. MEDIUM (2) covers recognisable conversational padding and AI-specific vocabulary. LOW (1) captures subtle indicators like tautological comments and generic boilerplate that require density to carry independent signal.

CRITICAL 9HIGH 179MEDIUM 2589LOW 12287

Directory Score Breakdown

This horizontal bar chart decomposes the repository's raw synthetic code score by top-level directory, allowing you to pinpoint precisely which modules or components carry the highest AI authorship density. Directories with disproportionately high scores relative to their size warrant targeted manual review: concentrated AI signatures often trace back to mass-generated configuration layers, auto-ported test suites, LLM-scaffolded boilerplate classes, or entire subsystems authored under heavy copilot assistance. Use this view to prioritise your human code-review effort.

Pattern Findings

The scanner identified 15064 distinct pattern matches across 26 syntactic categories. Each entry below represents a discrete location in the source code where the engine recorded a statistically significant AI authorship indicator. Expand any category row to inspect the individual file paths, line numbers, code snippets, and the lexical context (CODE, COMMENT, or STRING) in which each match was detected.

Reading the findings table: The Severity column indicates the diagnostic confidence level (CRITICAL / HIGH / MEDIUM / LOW). The Context column identifies whether the match occurred inside executable code, an inline comment, or a string literal — comment-context matches receive a ×1.5 weight because LLMs systematically over-annotate. The ⚡ bolt icon marks clustered matches: three or more patterns within a 10-line window, each receiving an additional ×1.5 density multiplier as dense clusters constitute far stronger evidence of synthetic authorship than isolated hits.

Decorative Section Separators2390 hits · 8426 pts

Severity	File	Line	Snippet	Context
MEDIUM	Cargo.toml	89	# ── Release profile — wheel size optimization ───────────────────────	COMMENT
MEDIUM	crates/headroom-core/config/pipeline.toml	35	# ─── Per-domain bloat estimator config ──────────────────────────────	COMMENT
MEDIUM	crates/headroom-core/config/pipeline.toml	79	# ─── Reformat configs ───────────────────────────────────────────────	COMMENT
MEDIUM	crates/headroom-core/config/pipeline.toml	106	# ─── JSON-array offload config (SmartCrusher wrapper) ─────────────	COMMENT
MEDIUM	crates/headroom-core/config/pipeline.toml	122	# ─── Diff-noise offload config ─────────────────────────────────────	COMMENT
MEDIUM	crates/headroom-core/tests/ccr_backends.rs	155	// ─── Redis-feature-gated tests ─────────────────────────────────────────	COMMENT
MEDIUM	crates/headroom-core/tests/auth_mode.rs	27	// ── Required matrix ──────────────────────────────────────────────	COMMENT
MEDIUM	crates/headroom-core/tests/auth_mode.rs	96	// ── Bonus matrix ──────────────────────────────────────────────────	COMMENT
MEDIUM	crates/headroom-core/tests/auth_mode.rs	152	// ── Performance ──────────────────────────────────────────────────	COMMENT
MEDIUM	crates/headroom-core/tests/live_zone_dispatch.rs	81	// ─── Routing tests ─────────────────────────────────────────────────────	COMMENT
MEDIUM	crates/headroom-core/tests/live_zone_dispatch.rs	343	// ─── Cache-safety invariant ────────────────────────────────────────────	COMMENT
MEDIUM	crates/headroom-core/tests/ccr_roundtrip.rs	226	// ─── PR8 additions: marker injection + walker unification ──────────	COMMENT
MEDIUM	crates/headroom-core/tests/ccr_roundtrip.rs	364	// ─── helpers ──────────────────────────────────────────────────────	COMMENT
MEDIUM	crates/headroom-core/tests/cache_control.rs	199	// ─── Property tests ───────────────────────────────────────────────────	COMMENT
MEDIUM	crates/headroom-core/benches/ccr_store.rs	26	// ─── Baseline: the old single-Mutex<HashMap> design ────────────────	COMMENT
MEDIUM	crates/headroom-core/src/auth_mode.rs	122	// ── User-Agent ───────────────────────────────────────────────	COMMENT
MEDIUM	crates/headroom-core/src/auth_mode.rs	147	// ── Authorization header ─────────────────────────────────────	COMMENT
MEDIUM	crates/headroom-core/src/auth_mode.rs	193	// ── Vendor-specific API-key headers ──────────────────────────	COMMENT
MEDIUM	crates/headroom-core/src/auth_mode.rs	204	// ── Default ──────────────────────────────────────────────────	COMMENT
MEDIUM	crates/headroom-core/src/transforms/unidiff_detector.rs	100	// ─── Tests ─────────────────────────────────────────────────────────────	COMMENT
MEDIUM	crates/headroom-core/src/transforms/diff_compressor.rs	47	// ─── Score-weight constants ────────────────────────────────────────────────	COMMENT
MEDIUM	crates/headroom-core/src/transforms/diff_compressor.rs	74	// ─── Public API ─────────────────────────────────────────────────────────────	COMMENT
MEDIUM	crates/headroom-core/src/transforms/diff_compressor.rs	530	// ─── Internal types ────────────────────────────────────────────────────────	COMMENT
MEDIUM	crates/headroom-core/src/transforms/diff_compressor.rs	580	// ─── Parser ────────────────────────────────────────────────────────────────	COMMENT
MEDIUM	crates/headroom-core/src/transforms/diff_compressor.rs	826	// ─── Scoring ───────────────────────────────────────────────────────────────	COMMENT
MEDIUM	crates/headroom-core/src/transforms/diff_compressor.rs	879	// ─── Hunk selection (max_hunks_per_file cap) ───────────────────────────────	COMMENT
MEDIUM	crates/headroom-core/src/transforms/diff_compressor.rs	962	// ─── Context trimming ──────────────────────────────────────────────────────	COMMENT
MEDIUM	crates/headroom-core/src/transforms/diff_compressor.rs	1045	// ─── Output formatter ──────────────────────────────────────────────────────	COMMENT
MEDIUM	crates/headroom-core/src/transforms/diff_compressor.rs	1115	// ─── Helpers ───────────────────────────────────────────────────────────────	COMMENT
MEDIUM	crates/headroom-core/src/transforms/diff_compressor.rs	1179	// ─── Tests ─────────────────────────────────────────────────────────────────	COMMENT
MEDIUM	crates/headroom-core/src/transforms/diff_compressor.rs	1277	// ─── Lossy-path tests ───────────────────────────────────────────────────	COMMENT
MEDIUM	crates/headroom-core/src/transforms/diff_compressor.rs	1477	// ─── Bug-fix tests (rename/combined-diff/no-newline/pre-diff) ──────────	COMMENT
MEDIUM⚡	crates/headroom-core/src/transforms/magika_detector.rs	487	// ── JSON ───────────────────────────────────────────────────	COMMENT
MEDIUM⚡	crates/headroom-core/src/transforms/magika_detector.rs	494	// ── Diffs ──────────────────────────────────────────────────	COMMENT
MEDIUM⚡	crates/headroom-core/src/transforms/magika_detector.rs	497	// ── HTML ───────────────────────────────────────────────────	COMMENT
MEDIUM⚡	crates/headroom-core/src/transforms/magika_detector.rs	500	// ── Source code ────────────────────────────────────────────	COMMENT
MEDIUM⚡	crates/headroom-core/src/transforms/magika_detector.rs	511	// ── Plain text-ish ─────────────────────────────────────────	COMMENT
MEDIUM⚡	crates/headroom-core/src/transforms/magika_detector.rs	519	// ── Default: passthrough ───────────────────────────────────	COMMENT
MEDIUM⚡	crates/headroom-core/src/transforms/magika_detector.rs	524	// ─── Tests ─────────────────────────────────────────────────────────────	COMMENT
MEDIUM	crates/headroom-core/src/transforms/live_zone.rs	112	// ─── Tunable constants (no magic numbers in the dispatch logic) ────────	COMMENT
MEDIUM	crates/headroom-core/src/transforms/live_zone.rs	140	// ─── Per-content-type byte thresholds ──────────────────────────────────	COMMENT
MEDIUM	crates/headroom-core/src/transforms/live_zone.rs	188	// ─── Public types ──────────────────────────────────────────────────────	COMMENT
MEDIUM	crates/headroom-core/src/transforms/live_zone.rs	524	// ─── Compressor singletons ─────────────────────────────────────────────	COMMENT
MEDIUM	crates/headroom-core/src/transforms/live_zone.rs	552	// ─── Public entry point ────────────────────────────────────────────────	COMMENT
MEDIUM	crates/headroom-core/src/transforms/live_zone.rs	808	// ─── Internal helpers ──────────────────────────────────────────────────	COMMENT
MEDIUM	crates/headroom-core/src/transforms/live_zone.rs	1730	// ─── Manifest accessor helpers (consumed by PyO3 binding) ─────────	COMMENT
MEDIUM	crates/headroom-core/src/transforms/live_zone.rs	1836	// ─── OpenAI Chat Completions live-zone dispatcher (Phase C PR-C2) ────────	COMMENT
MEDIUM	crates/headroom-core/src/transforms/live_zone.rs	2269	// ─── OpenAI Responses live-zone dispatcher (Phase C PR-C3) ────────────	COMMENT
MEDIUM	crates/headroom-core/src/transforms/detection.rs	63	// ── Tier 1: Magika ──────────────────────────────────────────	COMMENT
MEDIUM	crates/headroom-core/src/transforms/detection.rs	83	// ── Tier 2: unidiff parser ──────────────────────────────────	COMMENT
MEDIUM	crates/headroom-core/src/transforms/detection.rs	88	// ── Tier 3: fallthrough ─────────────────────────────────────	COMMENT
MEDIUM	crates/headroom-core/src/transforms/content_detector.rs	83	// ─── Regex patterns (compiled once, shared) ───────────────────────────	COMMENT
MEDIUM	crates/headroom-core/src/transforms/content_detector.rs	108	// ─── Code patterns by language ─────────────────────────────────────────	COMMENT
MEDIUM	crates/headroom-core/src/transforms/content_detector.rs	170	// ─── Log / build output patterns ───────────────────────────────────────	COMMENT
MEDIUM	crates/headroom-core/src/transforms/content_detector.rs	191	// ─── HTML patterns ─────────────────────────────────────────────────────	COMMENT
MEDIUM	crates/headroom-core/src/transforms/content_detector.rs	207	// ─── Public entry point ────────────────────────────────────────────────	COMMENT
MEDIUM	crates/headroom-core/src/transforms/content_detector.rs	272	// ─── Per-type detection helpers ────────────────────────────────────────	COMMENT
MEDIUM	crates/headroom-core/src/transforms/tag_protector.rs	738	// ─── Tiny byte-search helper ──────────────────────────────────────────	COMMENT
MEDIUM	crates/headroom-core/src/transforms/tag_protector.rs	945	// ─── Bug-fix tests (fixed_in_3e4) ─────────────────────────────────	COMMENT
MEDIUM	crates/headroom-core/src/transforms/tag_protector.rs	1017	// ─── Edge-case correctness ────────────────────────────────────────	COMMENT
2330 more matches not shown…

Hyper-Verbose Identifiers8138 hits · 7904 pts

Severity	File	Line	Snippet	Context
LOW	…ates/headroom-core/src/transforms/search_compressor.rs	13	//! src/models.py:15:class DataProcessor:	COMMENT
LOW⚡	plugins/headroom-oauth2/tests/test_oauth2.py	288	def test_install_noop_when_unset(monkeypatch):	CODE
LOW⚡	plugins/headroom-oauth2/tests/test_oauth2.py	298	def test_install_fail_closed_on_bad_config(monkeypatch):	CODE
LOW⚡	plugins/headroom-oauth2/tests/test_oauth2.py	305	def test_install_warns_for_envauth_backend(monkeypatch, caplog):	CODE
LOW	plugins/headroom-oauth2/tests/test_oauth2.py	142	def test_concurrent_single_flight(idp):	CODE
LOW	plugins/headroom-oauth2/tests/test_oauth2.py	156	def test_error_on_bad_status_hides_body(idp):	CODE
LOW	plugins/headroom-oauth2/tests/test_oauth2.py	171	def test_unreachable_token_url():	CODE
LOW	plugins/headroom-oauth2/tests/test_oauth2.py	219	def test_parse_headers_rejects_control_chars():	CODE
LOW	plugins/headroom-oauth2/tests/test_oauth2.py	225	def test_provider_from_env_wires_knobs(idp):	CODE
LOW	plugins/headroom-oauth2/tests/test_oauth2.py	242	def test_provider_from_env_none_when_unset():	CODE
LOW	plugins/headroom-oauth2/tests/test_oauth2.py	247	def test_middleware_injects_bearer(idp):	CODE
LOW	plugins/headroom-oauth2/tests/test_oauth2.py	258	def test_middleware_non_http_passthrough():	CODE
LOW	plugins/headroom-oauth2/tests/test_oauth2.py	322	def test_install_fail_closed_on_bad_timeout(monkeypatch):	CODE
LOW	plugins/headroom-oauth2/tests/test_oauth2.py	331	def test_parse_headers_rejects_bad_keys():	CODE
LOW	plugins/headroom-oauth2/tests/test_oauth2.py	343	def test_middleware_handles_missing_headers_key(idp):	CODE
LOW	plugins/headroom-oauth2/tests/test_oauth2.py	384	def test_allow_insecure_env_permits_nonloopback_http():	CODE
LOW	plugins/headroom-oauth2/tests/test_oauth2.py	397	def test_extra_params_cannot_override_canonical(idp):	CODE
LOW	plugins/headroom-oauth2/tests/test_oauth2.py	413	def test_auth_style_basic_via_env(idp):	CODE
LOW	plugins/headroom-oauth2/tests/test_oauth2.py	427	def test_scopes_comma_separated_via_env(idp):	CODE
LOW	plugins/headroom-oauth2/tests/test_oauth2.py	440	def test_expires_in_missing_falls_back(idp):	CODE
LOW	plugins/headroom-oauth2/tests/test_oauth2.py	447	def test_expires_in_non_numeric_falls_back(idp):	CODE
LOW	plugins/headroom-oauth2/tests/test_oauth2.py	463	def test_single_flight_on_refresh(idp):	CODE
LOW	plugins/headroom-oauth2/tests/test_oauth2.py	477	def test_install_sets_static_headers(monkeypatch):	CODE
LOW	plugins/openclaw/src/proxy-manager.ts	413	export function normalizeAndValidateProxyUrl(proxyUrl: string): string {	CODE
LOW	plugins/openclaw/src/gateway-config.ts	18	export function resolveGatewayProviderIds(config: Record<string, unknown> \| undefined): string[] {	CODE
LOW	plugins/openclaw/src/gateway-config.ts	31	function normalizeGatewayProviderIds(value: unknown): string[] {	CODE
LOW	plugins/openclaw/src/gateway-config.ts	67	export function applyGatewayProviderBaseUrlsInPlace(	CODE
LOW	plugins/openclaw/src/convert.ts	297	function normalizeAssistantMessage(message: Record<string, any>): Record<string, any> {	CODE
LOW	plugins/openclaw/src/convert.ts	321	function normalizeToolResultMessage(message: Record<string, any>): Record<string, any> {	CODE
LOW	plugins/openclaw/src/convert.ts	345	function normalizeAssistantContent(content: unknown): any[] {	CODE
LOW	plugins/openclaw/src/convert.ts	385	function normalizeToolResultContent(content: unknown): any[] {	CODE
LOW	plugins/openclaw/src/tools/headroom-retrieve.ts	15	export function createHeadroomRetrieveTool(config: RetrieveToolConfig) {	CODE
LOW	plugins/hermes/headroom_retrieve/__init__.py	47	def _handle_headroom_retrieve(args: dict, **kw) -> str:	CODE
LOW	plugins/opencode/src/retrieve.ts	18	export function createHeadroomRetrieveTool(config: RetrieveToolConfig) {	CODE
LOW	plugins/opencode/src/transport.ts	182	function normalizedOpenAiProxyPath(pathname: string): string \| undefined {	CODE
LOW	plugins/opencode/src/transport.ts	456	export function uninstallHeadroomTransport(): void {	CODE
LOW	plugins/opencode/src/provider.ts	73	export function buildOpencodeConfigContent(	CODE
LOW	plugins/opencode/src/provider.ts	85	export function buildOpencodeConfigContentJson(	CODE
LOW⚡	tests/test_quota_registry.py	88	def test_register_single_tracker():	CODE
LOW⚡	tests/test_quota_registry.py	95	def test_register_multiple_trackers():	CODE
LOW⚡	tests/test_quota_registry.py	104	def test_duplicate_key_raises():	CODE
LOW⚡	tests/test_quota_registry.py	111	def test_get_unknown_key_returns_none():	CODE
LOW⚡	tests/test_quota_registry.py	116	def test_trackers_property_is_snapshot():	CODE
LOW⚡	tests/test_quota_registry.py	131	def test_start_all_calls_available_trackers():	CODE
LOW⚡	tests/test_quota_registry.py	139	def test_start_all_skips_unavailable_trackers():	CODE
LOW⚡	tests/test_quota_registry.py	183	def test_passive_tracker_start_stop_are_noops():	CODE
LOW⚡	tests/test_quota_registry.py	197	def test_get_all_stats_includes_available_with_data():	CODE
LOW⚡	tests/test_quota_registry.py	205	def test_get_all_stats_excludes_unavailable():	CODE
LOW⚡	tests/test_quota_registry.py	212	def test_get_all_stats_excludes_none_data():	CODE
LOW⚡	tests/test_quota_registry.py	228	def test_get_stats_single_key():	CODE
LOW⚡	tests/test_quota_registry.py	234	def test_get_stats_missing_key_returns_none():	CODE
LOW⚡	tests/test_quota_registry.py	244	def test_get_quota_registry_returns_same_instance():	CODE
LOW⚡	tests/test_quota_registry.py	250	def test_reset_quota_registry_gives_fresh_instance():	CODE
LOW⚡	tests/test_quota_registry.py	259	def test_reset_quota_registry_clears_registrations():	CODE
LOW	tests/test_quota_registry.py	155	def test_stop_all_calls_stop_on_all_registered():	CODE
LOW	tests/test_quota_registry.py	165	def test_stop_all_continues_on_exception():	CODE
LOW	tests/test_sse_thinking_blocks.py	52	def test_thinking_delta_accumulated() -> None:	CODE
LOW	tests/test_sse_thinking_blocks.py	82	def test_signature_delta_preserved() -> None:	CODE
LOW	tests/test_sse_thinking_blocks.py	124	def test_citations_delta_accumulated() -> None:	CODE
LOW	tests/test_sse_thinking_blocks.py	174	def test_redacted_thinking_data_preserved() -> None:	CODE
8078 more matches not shown…

Unused Imports1718 hits · 1556 pts

Severity	File	Line	Context
LOW	docker/differential-network-capture/mitm_capture.py	3	CODE
LOW	plugins/headroom-oauth2/src/headroom_oauth2/provider.py	8	CODE
LOW	plugins/headroom-oauth2/src/headroom_oauth2/__init__.py	7	CODE
LOW	plugins/headroom-oauth2/src/headroom_oauth2/__init__.py	14	CODE
LOW	…gins/headroom-oauth2/src/headroom_oauth2/middleware.py	8	CODE
LOW	plugins/hermes/headroom_retrieve/__init__.py	13	CODE
LOW	tests/test_quota_registry.py	3	CODE
LOW	tests/test_sse_thinking_blocks.py	21	CODE
LOW	tests/test_paths_backward_compat.py	16	CODE
LOW	tests/test_provider_codex_install.py	1	CODE
LOW	tests/test_memory_decision.py	27	CODE
LOW	tests/e2e_cortex_savings.py	30	CODE
LOW	tests/e2e_cortex_savings.py	45	CODE
LOW	tests/test_openai_responses_t3_replay_regression.py	1	CODE
LOW	tests/test_h2_stream_reset_retry.py	11	CODE
LOW	tests/test_memory_wrapper.py	1	CODE
LOW	tests/test_codex_responses_passthrough_bytes.py	11	CODE
LOW	tests/test_tool_name_policy.py	1	CODE
LOW	tests/test_proxy_pipeline_lifecycle.py	1	CODE
LOW	tests/test_tool_injection_config.py	1	CODE
LOW	tests/test_codex_ws_per_frame_memory.py	1	CODE
LOW	tests/test_rtk_docker_availability.py	3	CODE
LOW	tests/test_memory_rank_policy.py	3	CODE
LOW	tests/test_hermes_passthrough_compression.py	15	CODE
LOW	tests/test_openai_responses_context_compaction.py	1	CODE
LOW	tests/test_perf_cli_filtering.py	10	CODE
LOW	tests/test_image_compression_offload.py	12	CODE
LOW	tests/test_turn_hooks.py	10	CODE
LOW	tests/test_proxy_loop_exception_health.py	3	CODE
LOW	tests/test_provider_registry_extended.py	1	CODE
LOW	tests/test_stats_new_input_savings_rate.py	10	CODE
LOW	tests/test_utils.py	1	CODE
LOW	tests/test_proxy_cache_ttl_metrics.py	3	CODE
LOW	tests/test_storage_backends.py	1	CODE
LOW	tests/test_provider_route_specs.py	1	CODE
LOW	tests/test_codex_rate_limits.py	3	CODE
LOW	tests/test_stateless_writers.py	8	CODE
LOW	tests/test_update_helpers.py	3	CODE
LOW	tests/test_stateless_toin.py	7	CODE
LOW	tests/test_proxy_per_provider_kompress.py	9	CODE
LOW	tests/test_proxy_byte_faithful_forwarding.py	19	CODE
LOW	tests/test_request_outcome.py	12	CODE
LOW	tests/test_proxy_cors.py	12	CODE
LOW	tests/test_subscription_tracker.py	1	CODE
LOW	tests/test_custom_base_passthrough_telemetry.py	1	CODE
LOW	tests/test_codex_ws_compression_scheduler.py	23	CODE
LOW	tests/test_proxy_savings_history.py	3	CODE
LOW	tests/test_proxy_anthropic_compression_diagnostics.py	16	CODE
LOW	tests/test_ccr_batch_store.py	1	CODE
LOW	tests/test_codex_client_stamp.py	9	CODE
LOW	tests/test_memory_tool_mode.py	14	CODE
LOW	tests/test_copilot_linux_secret.py	1	CODE
LOW	tests/test_fsutil.py	3	CODE
LOW	tests/test_ccr_response_handler_extra.py	1	CODE
LOW	tests/test_proxy_warmup.py	13	CODE
LOW	tests/test_tokenizers.py	3	CODE
LOW	tests/test_image_compression.py	7	CODE
LOW	tests/test_image_compression.py	26	CODE
LOW	tests/test_image_compression.py	222	CODE
LOW	tests/test_image_compression.py	223	CODE
1658 more matches not shown…

Excessive Try-Catch Wrapping738 hits · 670 pts

Severity	File	Line	Snippet	Context
LOW	claude_analysis_ttl.py	55	except Exception:	CODE
LOW	claude_analysis_ttl.py	65	except Exception:	CODE
LOW	claude_analysis_ttl.py	108	except Exception:	CODE
LOW	docker/differential-network-capture/mitm_capture.py	52	except Exception:	CODE
LOW	plugins/headroom-oauth2/src/headroom_oauth2/provider.py	133	except Exception:	CODE
LOW	plugins/headroom-oauth2/src/headroom_oauth2/__init__.py	119	except Exception as e: # pragma: no cover	CODE
LOW	tests/e2e_cortex_savings.py	382	except Exception as e:	CODE
LOW	tests/test_stateless_writers.py	32	except Exception:	CODE
LOW	tests/e2e_real_compression.py	86	except Exception:	CODE
LOW	tests/e2e_real_compression.py	124	except Exception:	CODE
MEDIUM	tests/test_compression_cache.py	477	def reader() -> None:	CODE
MEDIUM	tests/test_compression_cache.py	484	def writer() -> None:	CODE
LOW	tests/test_compression_cache.py	481	except Exception as e: # pragma: no cover	CODE
LOW	tests/test_compression_cache.py	489	except Exception as e: # pragma: no cover	CODE
LOW	tests/test_compression_cache.py	590	except Exception as e: # pragma: no cover	CODE
LOW	tests/test_compression_cache.py	630	except Exception as e: # pragma: no cover	CODE
LOW	tests/test_startup_log_noise.py	40	except Exception:	CODE
LOW	tests/test_startup_log_noise.py	60	except Exception:	CODE
MEDIUM	tests/test_compression_store.py	1056	def retrieve_item() -> None:	CODE
LOW	tests/test_compression_store.py	1036	except Exception as e:	CODE
LOW	tests/test_compression_store.py	1061	except Exception as e:	CODE
LOW	tests/test_compression_store.py	1096	except Exception as e:	CODE
LOW	tests/e2e_cortex_mcp.py	336	except Exception as e:	CODE
LOW	tests/e2e_cortex_quality.py	371	except Exception as exc:	CODE
MEDIUM	tests/test_memory_bridge.py	73	def wrapper(args, *kwargs):	CODE
LOW⚡	tests/test_memory_bridge.py	76	except Exception as exc:	CODE
LOW	tests/test_critical_gaps.py	364	except Exception as e:	CODE
LOW	tests/test_critical_gaps.py	476	except Exception as e:	CODE
LOW	tests/test_critical_gaps.py	490	except Exception as e:	CODE
LOW	tests/test_critical_gaps.py	501	except Exception as e:	CODE
MEDIUM	tests/test_memory_handler_concurrent_init.py	34	def wrapper(args, *kwargs):	CODE
LOW⚡	tests/test_memory_handler_concurrent_init.py	37	except Exception as exc:	CODE
LOW⚡	tests/test_memory_handler_concurrent_init.py	256	except Exception:	CODE
LOW⚡	tests/test_memory_handler_concurrent_init.py	259	except Exception: # pragma: no cover - env-dependent	CODE
LOW	tests/e2e_cortex_latency.py	434	except Exception as exc:	CODE
LOW	tests/e2e_cortex_proxy_mcp.py	107	except Exception:	CODE
LOW	tests/e2e_cortex_proxy_mcp.py	360	except Exception as e:	CODE
LOW⚡	tests/e2e_cortex_proxy.py	116	except Exception:	CODE
LOW	tests/e2e_cortex_proxy.py	217	except Exception as e:	CODE
LOW⚡	tests/e2e_cortex_proxy.py	274	except Exception as e:	CODE
LOW⚡	tests/e2e_cortex_proxy.py	288	except Exception as e:	CODE
LOW	tests/test_provider_codex_runtime.py	207	except Exception: # pragma: no cover - best effort poll	CODE
LOW	tests/e2e_ws_responses_compression.py	55	except Exception:	CODE
LOW	tests/test_ccr.py	351	except Exception as e:	CODE
MEDIUM	tests/test_ccr.py	336	def store_and_retrieve(i):	CODE
MEDIUM	tests/test_google_multimodal_e2e.py	59	print(f"Error: {response.text[:200]}")	CODE
MEDIUM	tests/test_google_multimodal_e2e.py	100	print(f"Error: {response.text[:200]}")	CODE
MEDIUM	tests/test_google_multimodal_e2e.py	159	print(f"Error: {response.text[:200]}")	CODE
MEDIUM	tests/test_google_multimodal_e2e.py	223	print(f"Error: {response.text[:300]}")	CODE
MEDIUM	tests/test_google_multimodal_e2e.py	268	print(f"Error: {response.text[:200]}")	CODE
MEDIUM	tests/test_google_multimodal_e2e.py	315	print(f"Error: {response.text[:300]}")	CODE
LOW	tests/e2e_ws_codex_usage_headers.py	72	except Exception:	CODE
LOW	tests/e2e_ws_codex_usage_headers.py	220	except Exception as exc: # noqa: BLE001 - best-effort secondary check	CODE
MEDIUM	tests/e2e_ws_codex_usage_headers.py	88	def _handler(self, ws):	CODE
LOW	tests/test_proxy_streaming_resilience.py	376	except Exception as e:	CODE
MEDIUM	tests/test_cache/test_backends.py	263	def setter() -> None:	CODE
MEDIUM	tests/test_cache/test_backends.py	270	def getter() -> None:	CODE
MEDIUM	tests/test_cache/test_backends.py	277	def deleter() -> None:	CODE
LOW⚡	tests/test_cache/test_backends.py	267	except Exception as e:	CODE
LOW⚡	tests/test_cache/test_backends.py	274	except Exception as e:	CODE
678 more matches not shown…

Over-Commented Block725 hits · 666 pts

Severity	File	Line	Snippet	Context
LOW	Cargo.toml	41	# I1 (byte-faithful passthrough on unmutated bytes; see REALIGNMENT/02-	COMMENT
LOW	Cargo.toml	61	axum = "0.7"	COMMENT
LOW	Cargo.toml	81	# (ADC) → bearer token for the `Authorization: Bearer <token>` header.	COMMENT
LOW	Cargo.toml	101	# * Single codegen unit (better inlining + dead-code elim, at the	COMMENT
LOW	pyproject.toml	101	# ML-based compression with Kompress (ModernBERT).	COMMENT
LOW	pyproject.toml	141	"sentence-transformers>=2.2.0; sys_platform == 'darwin'",	COMMENT
LOW	pyproject.toml	161	# * rapidocr 3.x — engine-agnostic core, supports Python 3.13+.	COMMENT
LOW	pyproject.toml	281	"respx>=0.20.0", # HTTP mock transport for passthrough handler tests	COMMENT
LOW	pyproject.toml	301	# - ml (torch — the PyTorch Kompress backend; ONNX path in [proxy] still	COMMENT
LOW	pyproject.toml	321	Documentation = "https://headroom-docs.vercel.app/docs"	COMMENT
LOW	pyproject.toml	361	# `pypi.netflix.net/simple` — and bakes those URLs into uv.lock, which	COMMENT
LOW	.gitguardian.yaml	1	# GitGuardian configuration — secret-detection allowlist	COMMENT
LOW	rust-toolchain.toml	1	[toolchain]	COMMENT
LOW	claude_analysis_ttl.py	181	f" Median post-idle write (5-60min gap) : {med_5_60:>10,} tokens "	COMMENT
LOW	RUST_DEV.md	101	# Top compressors by invocation count (last process lifetime)	COMMENT
LOW	crates/headroom-parity/src/lib.rs	421	"was_modified": result.was_modified,	COMMENT
LOW	crates/headroom-core/Cargo.toml	21	# `hf-hub` is the HuggingFace Hub client. We use the blocking `ureq` transport	COMMENT
LOW	crates/headroom-core/Cargo.toml	41	# sizing pipeline uses `zlib.compress(..., level=1)` to validate the chosen	COMMENT
LOW	crates/headroom-core/Cargo.toml	61	# (release matrix, e2e/wrap, e2e/init, devcontainer, ci.yml) needs	COMMENT
LOW	crates/headroom-core/Cargo.toml	81	# in a line in O(n + m) — orders of magnitude faster than running N	COMMENT
LOW	crates/headroom-core/Cargo.toml	101	# `<<ccr:HASH>>` marker. Pinning the algorithm + truncation length here	COMMENT
LOW	…core/proptest-regressions/transforms/tag_protector.txt	1	# Seeds for failure cases proptest has generated in the past. It is	COMMENT
LOW	crates/headroom-core/config/pipeline.toml	1	# Compression pipeline default configuration.	COMMENT
LOW	crates/headroom-core/config/pipeline.toml	21		COMMENT
LOW	crates/headroom-core/config/pipeline.toml	41		COMMENT
LOW	crates/headroom-core/config/pipeline.toml	81	# Reformats pack input denser without losing information. Configs here	COMMENT
LOW	crates/headroom-core/config/pipeline.toml	101	# Floor on the number of constant tokens a template must have. A	COMMENT
LOW	crates/headroom-core/config/pipeline.toml	121		COMMENT
LOW	crates/headroom-core/tests/recommendations_loader.rs	1	//! Integration tests for `transforms::recommendations` (PR-B5).	COMMENT
LOW	crates/headroom-core/tests/auth_mode.rs	1	//! Integration tests for `headroom_core::auth_mode::classify`.	COMMENT
LOW	…ates/headroom-core/tests/live_zone_token_validation.rs	61	let s = std::str::from_utf8(bytes).expect("body is UTF-8 JSON");	COMMENT
LOW	…ates/headroom-core/tests/live_zone_token_validation.rs	81	// tokenize into many short BPE pieces post-rewrite.	COMMENT
LOW	…ates/headroom-core/tests/live_zone_token_validation.rs	201	proptest! {	COMMENT
LOW	crates/headroom-core/tests/tokenizer_proptest.rs	41	let est = EstimatingCounter::default();	COMMENT
LOW	crates/headroom-core/tests/live_zone_dispatch.rs	1	//! Integration tests for the PR-B3 live-zone dispatcher.	COMMENT
LOW	crates/headroom-core/tests/cache_control.rs	1	//! Unit + property tests for the `cache_control` walker (PR-A4).	COMMENT
LOW	crates/headroom-core/tests/cache_control.rs	61	});	COMMENT
LOW	crates/headroom-core/tests/cache_control.rs	201	proptest! {	COMMENT
LOW	crates/headroom-core/benches/auth_mode.rs	1	//! Criterion benchmark for the auth-mode classifier (Phase F PR-F1).	COMMENT
LOW	crates/headroom-core/benches/ccr_store.rs	1	//! CCR store throughput benchmark — single-threaded and multi-threaded.	COMMENT
LOW	crates/headroom-core/src/compression_policy.rs	1	//! Per-auth-mode compression policy — Phase F PR-F2.1, extended in F2.2.	COMMENT
LOW	crates/headroom-core/src/compression_policy.rs	21	//! 2. Test surface. `for_mode(AuthMode) -> CompressionPolicy` is	COMMENT
LOW	crates/headroom-core/src/compression_policy.rs	41	//! `CacheAligner` transform's `should_apply` MUST return `False`.	COMMENT
LOW	crates/headroom-core/src/compression_policy.rs	61	//!	COMMENT
LOW	crates/headroom-core/src/compression_policy.rs	81	//!	COMMENT
LOW	crates/headroom-core/src/compression_policy.rs	101	use crate::auth_mode::AuthMode;	COMMENT
LOW	crates/headroom-core/src/compression_policy.rs	121	/// Subscription: conservative — flag volatile content earlier (32	COMMENT
LOW	crates/headroom-core/src/compression_policy.rs	141	pub const CACHE_READ_MULTIPLIER: f32 = 0.1;	COMMENT
LOW	crates/headroom-core/src/compression_policy.rs	161	///	COMMENT
LOW	crates/headroom-core/src/compression_policy.rs	241	///	COMMENT
LOW	crates/headroom-core/src/compression_policy.rs	301	) -> bool {	COMMENT
LOW	crates/headroom-core/src/auth_mode.rs	1	//! Auth-mode classifier — Phase F PR-F1.	COMMENT
LOW	crates/headroom-core/src/auth_mode.rs	21	//!	COMMENT
LOW	crates/headroom-core/src/auth_mode.rs	61	AuthMode::OAuth => "oauth",	COMMENT
LOW	crates/headroom-core/src/auth_mode.rs	81	"codex-cli/",	COMMENT
LOW	crates/headroom-core/src/auth_mode.rs	101	/// [`AuthMode::Payg`] (Anthropic / OpenAI API key).	COMMENT
LOW	crates/headroom-core/src/auth_mode.rs	181	// Unknown bearer shape — fall through to header-based	COMMENT
LOW	crates/headroom-core/src/onnx_cpu.rs	1	//! Shared CPU-capability guard for the precompiled ONNX Runtime binary.	COMMENT
LOW	crates/headroom-core/src/cache_control.rs	1	//! Customer `cache_control` marker walker for Anthropic `/v1/messages`	COMMENT
LOW	crates/headroom-core/src/cache_control.rs	21	//! For an Anthropic request body parsed into `serde_json::Value`,	COMMENT
665 more matches not shown…

Cross-Language Confusion74 hits · 368 pts

Severity	File	Line	Snippet	Context
HIGH	tests/test_proxy_savings_history.py	1334	assert "if (this.historySelectedSeriesKey === 'history') return null;" in html	CODE
HIGH⚡	tests/test_cortex_code_compression.py	13	cd headroom && uv run python tests/test_cortex_code_compression.py	STRING
HIGH⚡	tests/test_cortex_code_compression.py	16	cd headroom && uv run --with pytest pytest tests/test_cortex_code_compression.py -v -s	STRING
HIGH⚡	tests/test_observed_wire_shapes.py	305	"cd /x && cat a.py",	CODE
HIGH⚡	tests/test_observed_wire_shapes.py	306	"cd /x && cat -A a.py \| head -60",	CODE
HIGH⚡	tests/test_observed_wire_shapes.py	358	"cd /tmp/core && rg -l safe_math --type py",	CODE
HIGH⚡	tests/test_observed_wire_shapes.py	359	"cd /tmp/core && grep -rn foo suma/",	CODE
HIGH⚡	tests/test_observed_wire_shapes.py	361	"cd /a && cd b && rg pat", # chained cds	CODE
HIGH⚡	tests/test_observed_wire_shapes.py	367	for cmd in ["cd /x && cat a.py", "cd /x && python -c 'x'", "cd /x && ls -la"]:	CODE
HIGH⚡	tests/test_observed_wire_shapes.py	372	assert _stripcd("cd /tmp/core && rg x") == "rg x"	CODE
HIGH⚡	tests/test_observed_wire_shapes.py	374	assert _stripcd("cd a && cd b && grep x") == "grep x"	CODE
HIGH⚡	tests/test_observed_wire_shapes.py	480	anthropic_input = {"command": "cd /tmp/core && cat suma/x.py"} # Anthropic: dict	CODE
HIGH⚡	tests/test_observed_wire_shapes.py	481	openai_args = json.dumps({"command": "cd /tmp/core && cat suma/x.py"}) # OpenAI: JSON string	CODE
HIGH⚡	tests/test_observed_wire_shapes.py	486	assert not _isread2(_cmdtext({"command": "cd /x && rg pat"}))	CODE
HIGH	tests/test_observed_wire_shapes.py	51	"input": {"command": 'cd /tmp/core && rg -l "safe_math" --type py \| head'},	CODE
HIGH	tests/test_observed_wire_shapes.py	394	"content": "THOUGHT: look\n```bash\ncd /r && cat x.py\n```",	CODE
HIGH	tests/test_observed_wire_shapes.py	423	_fenced_shell_command("T\n```mswea_bash_command\ncd /r && cat x.py\n```")	CODE
HIGH	tests/test_observed_wire_shapes.py	424	== "cd /r && cat x.py"	CODE
HIGH	tests/test_observed_wire_shapes.py	435	"content": "T\n```mswea_bash_command\ncd /r && cat a.py\n```",	CODE
HIGH	tests/test_observed_wire_shapes.py	444	"content": "T\n```mswea_bash_command\ncd /r && grep -rn foo .\n```",	CODE
HIGH	tests/test_observed_wire_shapes.py	557	{"role": "assistant", "content": "", "tool_calls": [tc("c_read", "cd /r && cat a.py")]},	CODE
HIGH	tests/test_observed_wire_shapes.py	562	"tool_calls": [tc("c_grep", "cd /r && grep -rn foo .")],	CODE
HIGH	tests/test_toin_feedback.py	413	'data: {"type":"message_start","message":{"id":"msg_01","model":"claude-3-5-sonnet-20241022","role":"assista	CODE
HIGH	tests/test_openai_codex_ws_timings.py	230	with ``upstream_first_event`` absent (null)."""	STRING
HIGH	tests/test_code_aware_brace_comment_regressions.py	61	this.users.push(user);	CODE
HIGH	tests/test_proxy_streaming_request_logger.py	260	b'"content":[],"stop_reason":null,"usage":{'	CODE
HIGH	tests/test_cache_prefix_overlay.py	112	cmd = "cd /tmp/core && cat suma/apps/underwriting/followup/service.py"	CODE
HIGH⚡	tests/test_cache_prefix_overlay.py	138	_openai_asst("cd /tmp/core && rg -n foo ."),	CODE
HIGH⚡	tests/test_cache_prefix_overlay.py	140	_openai_asst("cd /tmp/core && cat foo.py"), # NEW delta (assistant)	CODE
HIGH	…sts/test_integrations/langchain/test_langchain_live.py	6	set -a && source .env && set +a && pytest tests/test_integrations/langchain/test_langchain_live.py -v -s	STRING
HIGH	tests/test_memory/test_budget.py	83	"Use source .venv/bin/activate && pytest for running tests", importance=0.5	CODE
HIGH	tests/test_memory/test_budget.py	86	"Use source .venv/bin/activate && pytest for running tests", importance=0.8	CODE
HIGH⚡	tests/test_memory/test_traffic_learner.py	199	assert _bash_first_binary("source .venv/bin/activate && pytest -x") == "pytest"	CODE
HIGH⚡	tests/test_memory/test_traffic_learner.py	303	tool_input={"command": "source /project/.venv/bin/activate && pytest"},	CODE
HIGH⚡	tests/test_memory/test_traffic_learner.py	1672	assert _normalize_bash_for_hash("cd /tmp && ls") == "cd /tmp"	CODE
HIGH	tests/test_memory/test_traffic_learner.py	269	tool_input={"command": "source .venv/bin/activate && ruff check ."},	CODE
HIGH	tests/parity/recorder.py	667	"type Maybe<T> = T \| null;\n"	CODE
HIGH	tests/test_compression/test_json_handler.py	78	content = '{"value": null}'	CODE
HIGH	tests/test_compression/test_json_handler.py	333	content = '{"a": true, "b": false, "c": null}'	CODE
HIGH	tests/test_compression/test_json_handler.py	376	content = '{"str": "hello", "num": 1.5, "bool": true, "null": null}'	CODE
HIGH	tests/test_compression/test_evals.py	438	private socket: Socket \| null = null;	STRING
HIGH	tests/test_compression/test_evals.py	444	private heartbeatTimer: NodeJS.Timer \| null = null;	STRING
HIGH	tests/test_compression/test_evals.py	507	this.socket = null;	STRING
HIGH	tests/test_compression/test_evals.py	542	this.pendingOperations.push(operation);	STRING
HIGH	tests/test_compression/test_evals.py	638	this.heartbeatTimer = null;	STRING
HIGH	tests/test_compression/test_evals.py	643	if (this.pendingOperations.length > 0 && this.socket?.connected) {	STRING
HIGH	benchmarks/adversarial_ccr_tests.py	565	'[{"id": 1}, null, {"id": 2}]', # Null in array	CODE
HIGH	benchmarks/prefix_cache_benchmark.py	701	print(" source .env && python benchmarks/prefix_cache_benchmark.py")	CODE
HIGH⚡	benchmarks/prefix_cache_benchmark.py	16	source .env && python benchmarks/prefix_cache_benchmark.py	STRING
HIGH⚡	benchmarks/prefix_cache_benchmark.py	19	source .env && python benchmarks/prefix_cache_benchmark.py --turns 5	STRING
HIGH⚡	benchmarks/prefix_cache_benchmark.py	22	source .env && python benchmarks/prefix_cache_benchmark.py --model claude-sonnet-4-6	STRING
HIGH⚡	benchmarks/comprehensive_eval.py	263	data: null,	STRING
HIGH⚡	benchmarks/comprehensive_eval.py	264	error: null,	STRING
HIGH⚡	benchmarks/comprehensive_eval.py	274	setState(prev => ({ ...prev, loading: true, error: null }));	STRING
HIGH⚡	benchmarks/comprehensive_eval.py	278	if (mountedRef.current && callId === lastCallId.current) {	STRING
HIGH⚡	benchmarks/comprehensive_eval.py	279	setState({ data: result, error: null, loading: false });	STRING
HIGH⚡	benchmarks/comprehensive_eval.py	284	if (mountedRef.current && callId === lastCallId.current) {	STRING
HIGH⚡	benchmarks/comprehensive_eval.py	286	setState({ data: null, error: err, loading: false });	STRING
HIGH	benchmarks/comprehensive_eval.py	246	data: T \| null;	STRING
HIGH	benchmarks/comprehensive_eval.py	247	error: Error \| null;	STRING
14 more matches not shown…

Deep Nesting432 hits · 360 pts

Severity	File	Line	Context
LOW	claude_analysis_ttl.py	45	CODE
LOW	claude_analysis_ttl.py	98	CODE
LOW	plugins/headroom-oauth2/src/headroom_oauth2/__init__.py	95	CODE
LOW	tests/test_output_only_request_blocks.py	42	CODE
LOW	tests/e2e_real_compression.py	79	CODE
LOW	tests/e2e_real_compression.py	252	CODE
LOW	tests/test_read_maturation_handler_nobust.py	135	CODE
LOW	tests/test_handler_outcome_tag_invariant.py	38	CODE
LOW	tests/test_lean_ctx_installer.py	115	CODE
LOW	tests/test_lean_ctx_installer.py	159	CODE
LOW	tests/test_proxy_system_prompt_immutable.py	371	CODE
LOW	tests/test_memory_integration.py	280	CODE
LOW	tests/test_critical_gaps.py	163	CODE
LOW	tests/test_critical_gaps.py	340	CODE
LOW	tests/test_critical_gaps.py	355	CODE
LOW	tests/test_realignment_live_multi_turn.py	558	CODE
LOW	tests/test_realignment_live_multi_turn.py	861	CODE
LOW	tests/test_tag_protector_invariant.py	121	CODE
LOW	tests/_mcp_stub.py	58	CODE
LOW	tests/e2e_cortex_proxy_mcp.py	148	CODE
LOW	tests/test_proxy_anthropic_cache_stability.py	1351	CODE
LOW	tests/test_anthropic_pre_upstream_backpressure.py	877	CODE
LOW	tests/test_anthropic_pre_upstream_backpressure.py	882	CODE
LOW	tests/test_tool_result_interceptors.py	28	CODE
LOW	tests/e2e_ws_responses_compression.py	48	CODE
LOW	tests/test_claude_session_mode_benchmark.py	429	CODE
LOW	tests/test_claude_session_mode_benchmark.py	437	CODE
LOW	tests/test_release_workflows.py	354	CODE
LOW	tests/e2e_ws_codex_usage_headers.py	65	CODE
LOW	tests/test_cache/test_client_integration.py	38	CODE
LOW	tests/test_integrations/langchain/test_chat_model.py	1041	CODE
LOW	tests/test_integrations/agno/test_model.py	1042	CODE
LOW	tests/test_cli/test_wrap_copilot.py	47	CODE
LOW	tests/test_cli/test_wrap_copilot.py	245	CODE
LOW	tests/test_cli/test_wrap_codex.py	1518	CODE
LOW	tests/test_cli/test_wrap_codex.py	1673	CODE
LOW	tests/test_cli/test_wrap_codex.py	1730	CODE
LOW	tests/test_cli/test_main_help_version.py	50	CODE
LOW	tests/integrations/test_strands/test_hooks.py	161	CODE
LOW	tests/test_transforms/test_code_compressor.py	1334	CODE
LOW	tests/test_transforms/test_read_lifecycle.py	619	CODE
LOW	examples/strands_mcp_dispatch_test.py	66	CODE
LOW	examples/strands_mcp_dispatch_test.py	122	CODE
LOW	examples/strands_bundle_demo.py	76	CODE
LOW	examples/strands_via_proxy_demo.py	122	CODE
LOW	examples/strands_via_proxy_demo.py	275	CODE
LOW	examples/langchain_demo/run_comparison.py	154	CODE
LOW	examples/langchain_demo/run_comparison.py	241	CODE
LOW	examples/langchain_demo/show_compression.py	35	CODE
LOW	benchmarks/conftest.py	52	CODE
LOW	benchmarks/adversarial_ccr_tests.py	1177	CODE
LOW	benchmarks/adversarial_ccr_tests.py	1582	CODE
LOW	benchmarks/adversarial_ccr_tests.py	1765	CODE
LOW	benchmarks/adversarial_ccr_tests.py	1838	CODE
LOW	benchmarks/adversarial_ccr_tests.py	1787	CODE
LOW	benchmarks/agent_cost_benchmark.py	133	CODE
LOW	benchmarks/agent_cost_benchmark.py	352	CODE
LOW	benchmarks/agent_cost_benchmark.py	421	CODE
LOW	benchmarks/agent_cost_benchmark.py	597	CODE
LOW	benchmarks/cache_validation_bundle.py	49	CODE
372 more matches not shown…

Modern Structural Boilerplate266 hits · 268 pts

Severity	File	Line	Snippet	Context
LOW	plugins/headroom-oauth2/src/headroom_oauth2/__init__.py	16	__all__ = ["install", "OAuth2ClientCredentials", "OAuth2Error", "OAuth2Middleware", "parse_headers"]	CODE
LOW⚡	tests/test_providers_opencode_config.py	22	def _set_test_home(monkeypatch: pytest.MonkeyPatch, tmp_path: Path) -> None:	CODE
LOW	tests/test_cli_learn.py	405	def set_context_target(self, target: str \| None) -> None:	CODE
LOW⚡	tests/test_ssl_context.py	82	def set_alpn_protocols(self, protocols: list[str]) -> None:	CODE
LOW	…ts/test_proxy/test_anthropic_ccr_deferred_injection.py	64	def update_from_result(self, originals, compressed) -> None: # noqa: ARG002	CODE
LOW	tests/parity/recorder.py	845	__all__ = [	CODE
LOW	tests/test_cli/test_wrap_codex.py	28	def _set_test_home(monkeypatch: pytest.MonkeyPatch, tmp_path: Path) -> None:	CODE
LOW⚡	tests/test_cli/test_wrap_opencode.py	21	def _set_test_home(monkeypatch: pytest.MonkeyPatch, tmp_path: Path) -> None:	CODE
LOW	tests/test_cli/test_wrap_bridge.py	24	def _set_test_home(monkeypatch, tmp_path: Path) -> None:	CODE
LOW	tests/test_compression/test_evals.py	132	logger = logging.getLogger(__name__)	STRING
LOW	benchmarks/__init__.py	36	__all__ = [	CODE
LOW	benchmarks/scenarios/__init__.py	22	__all__ = [	CODE
LOW⚡	scripts/version-sync.py	40	def update_package_json(file_path: Path, version: str) -> None:	CODE
LOW⚡	scripts/version-sync.py	50	def update_plugin_manifest(file_path: Path, version: str) -> None:	CODE
LOW⚡	scripts/version-sync.py	60	def update_marketplace_manifest(file_path: Path, version: str) -> None:	CODE
LOW	scripts/version-sync.py	77	def update_plugin_versions(root: Path, version: str) -> None:	CODE
LOW	scripts/version-sync.py	90	def update_openclaw_package_json(file_path: Path, version: str, sdk_version: str) -> None:	CODE
LOW	scripts/version-sync.py	102	def update_pyproject_version(root: Path, version: str) -> None:	CODE
LOW	e2e/_lib/harness.py	327	__all__ = [	CODE
LOW	e2e/_lib/__init__.py	23	__all__ = [	CODE
LOW⚡	headroom/paths.py	108	def set_process_stateless(value: bool) -> None:	CODE
LOW	headroom/paths.py	408	__all__ = [	CODE
LOW	headroom/binaries.py	42	logger = logging.getLogger(__name__)	CODE
LOW	headroom/binaries.py	44	__all__ = [	CODE
LOW	headroom/update_check.py	31	logger = logging.getLogger(__name__)	CODE
LOW	headroom/update_check.py	297	__all__ = [	CODE
LOW	headroom/savings_ledger.py	393	__all__ = [	CODE
LOW	headroom/client.py	38	logger = logging.getLogger(__name__)	CODE
LOW	headroom/__init__.py	91	__all__ = [	CODE
LOW	headroom/copilot_linux_secret.py	14	logger = logging.getLogger(__name__)	CODE
LOW	headroom/shared_context.py	33	logger = logging.getLogger(__name__)	CODE
LOW	headroom/compress.py	69	logger = logging.getLogger(__name__)	CODE
LOW	headroom/_ort.py	54	logger = logging.getLogger(__name__)	CODE
LOW	headroom/copilot_auth.py	26	logger = logging.getLogger(__name__)	CODE
LOW	headroom/copilot_auth.py	733	def _set_header_default(headers: dict[str, str], name: str, value: str) -> None:	CODE
LOW	headroom/agent_savings.py	11	logger = logging.getLogger(__name__)	CODE
LOW	headroom/copilot_macos_keychain.py	14	logger = logging.getLogger(__name__)	CODE
LOW	headroom/mcp_registry/codex.py	29	logger = logging.getLogger(__name__)	CODE
LOW	headroom/mcp_registry/claude.py	26	logger = logging.getLogger(__name__)	CODE
LOW	headroom/mcp_registry/__init__.py	30	__all__ = [	CODE
LOW	headroom/mcp_registry/opencode.py	18	logger = logging.getLogger(__name__)	CODE
LOW	headroom/relevance/embedding.py	58	logger = logging.getLogger(__name__)	CODE
LOW	headroom/relevance/__init__.py	57	__all__ = [	CODE
LOW	headroom/tokenizers/mistral.py	17	logger = logging.getLogger(__name__)	CODE
LOW	headroom/tokenizers/registry.py	20	logger = logging.getLogger(__name__)	CODE
LOW	headroom/tokenizers/__init__.py	58	__all__ = [	CODE
LOW	headroom/tokenizers/huggingface.py	17	logger = logging.getLogger(__name__)	CODE
LOW	headroom/tokenizers/tiktoken_counter.py	21	logger = logging.getLogger(__name__)	CODE
LOW	headroom/reporting/__init__.py	5	__all__ = ["generate_report"]	CODE
LOW	headroom/capture/__init__.py	11	__all__ = [	CODE
LOW	headroom/capture/network_diff.py	20	logger = logging.getLogger(__name__)	CODE
LOW	headroom/evals/suite_runner.py	29	logger = logging.getLogger(__name__)	CODE
LOW	headroom/evals/prompt_comparison.py	35	logger = logging.getLogger(__name__)	CODE
LOW	headroom/evals/cost_tracker.py	13	logger = logging.getLogger(__name__)	CODE
LOW	headroom/evals/__init__.py	83	__all__ = [	CODE
LOW⚡	headroom/evals/html_oss_benchmarks.py	32	logger = logging.getLogger(__name__)	CODE
LOW	headroom/evals/comprehensive_benchmark.py	42	logger = logging.getLogger(__name__)	CODE
LOW⚡	headroom/evals/batch_compression_eval.py	51	logger = logging.getLogger(__name__)	CODE
LOW	headroom/evals/html_extraction.py	21	logger = logging.getLogger(__name__)	CODE
LOW	headroom/evals/memory/runner.py	34	logger = logging.getLogger(__name__)	CODE
206 more matches not shown…

Cross-File Repetition50 hits · 250 pts

Severity	File	Snippet	Context
HIGH	tests/test_integrations/langchain/test_streaming.py	tests for behavior when langchain is not available.	STRING
HIGH	tests/test_integrations/langchain/test_memory.py	tests for behavior when langchain is not available.	STRING
HIGH	tests/test_integrations/langchain/test_retriever.py	tests for behavior when langchain is not available.	STRING
HIGH	tests/test_integrations/langchain/test_agents.py	tests for behavior when langchain is not available.	STRING
HIGH	tests/test_integrations/langchain/test_streaming.py	_check_langchain_available raises importerror when not available.	STRING
HIGH	tests/test_integrations/langchain/test_memory.py	_check_langchain_available raises importerror when not available.	STRING
HIGH	tests/test_integrations/langchain/test_retriever.py	_check_langchain_available raises importerror when not available.	STRING
HIGH	tests/test_integrations/langchain/test_agents.py	_check_langchain_available raises importerror when not available.	STRING
HIGH	tests/test_cli/test_wrap_goose.py	if the openhands binary is missing the command must fail with a clear error.	STRING
HIGH	tests/test_cli/test_wrap_opencode.py	if the openhands binary is missing the command must fail with a clear error.	STRING
HIGH	tests/test_cli/test_wrap_openhands.py	if the openhands binary is missing the command must fail with a clear error.	STRING
HIGH	headroom/memory/system.py	retrieve a specific memory by id. args: memory_id: the memory identifier. returns: memory if found, none otherwise.	STRING
HIGH	headroom/memory/backends/mem0_system_adapter.py	retrieve a specific memory by id. args: memory_id: the memory identifier. returns: memory if found, none otherwise.	STRING
HIGH	headroom/memory/backends/direct_mem0.py	retrieve a specific memory by id. args: memory_id: the memory identifier. returns: memory if found, none otherwise.	STRING
HIGH	headroom/memory/ports.py	get a memory by id. args: memory_id: the unique identifier of the memory. returns: the memory if found, none otherwise.	STRING
HIGH	headroom/memory/adapters/sqlite.py	get a memory by id. args: memory_id: the unique identifier of the memory. returns: the memory if found, none otherwise.	STRING
HIGH	headroom/memory/backends/local.py	get a memory by id. args: memory_id: the unique identifier of the memory. returns: the memory if found, none otherwise.	STRING
HIGH	headroom/memory/ports.py	get the full history chain for a memory. follows supersession links to return all versions of a memory. args: memory_id:	STRING
HIGH	headroom/memory/adapters/sqlite.py	get the full history chain for a memory. follows supersession links to return all versions of a memory. args: memory_id:	STRING
HIGH	headroom/memory/core.py	get the full history chain for a memory. follows supersession links to return all versions of a memory. args: memory_id:	STRING
HIGH	headroom/memory/ports.py	clear all memories at or below a scope level. args: user_id: required user scope. session_id: if provided, clear session	STRING
HIGH	headroom/memory/core.py	clear all memories at or below a scope level. args: user_id: required user scope. session_id: if provided, clear session	STRING
HIGH	headroom/memory/adapters/sqlite.py	clear all memories at or below a scope level. args: user_id: required user scope. session_id: if provided, clear session	STRING
HIGH	headroom/memory/ports.py	remove multiple memories from the text index (protocol-compliant). args: memory_ids: list of memory ids to remove. retur	STRING
HIGH	headroom/memory/adapters/hnsw.py	remove multiple memories from the text index (protocol-compliant). args: memory_ids: list of memory ids to remove. retur	STRING
HIGH	headroom/memory/adapters/fts5.py	remove multiple memories from the text index (protocol-compliant). args: memory_ids: list of memory ids to remove. retur	STRING
HIGH	headroom/memory/adapters/sqlite_vector.py	remove multiple memories from the text index (protocol-compliant). args: memory_ids: list of memory ids to remove. retur	STRING
HIGH	headroom/memory/ports.py	retrieve an entity by id. args: entity_id: the unique identifier of the entity. returns: the entity if found, none other	STRING
HIGH	headroom/memory/adapters/sqlite_graph.py	retrieve an entity by id. args: entity_id: the unique identifier of the entity. returns: the entity if found, none other	STRING
HIGH	headroom/memory/adapters/graph.py	retrieve an entity by id. args: entity_id: the unique identifier of the entity. returns: the entity if found, none other	STRING
HIGH	headroom/memory/ports.py	delete a single relationship. args: relationship_id: the unique identifier of the relationship. returns: true if the rel	STRING
HIGH	headroom/memory/adapters/sqlite_graph.py	delete a single relationship. args: relationship_id: the unique identifier of the relationship. returns: true if the rel	STRING
HIGH	headroom/memory/adapters/graph.py	delete a single relationship. args: relationship_id: the unique identifier of the relationship. returns: true if the rel	STRING
HIGH	headroom/memory/adapters/fts5.py	get a new database connection (thread-safe pattern). returns: a new sqlite connection with row factory configured.	STRING
HIGH	headroom/memory/adapters/sqlite_graph.py	get a new database connection (thread-safe pattern). returns: a new sqlite connection with row factory configured.	STRING
HIGH	headroom/memory/adapters/sqlite.py	get a new database connection (thread-safe pattern). returns: a new sqlite connection with row factory configured.	STRING
HIGH	headroom/memory/adapters/graph.py	get memory statistics for the memorytracker. returns: componentstats with current memory usage.	STRING
HIGH	headroom/memory/adapters/hnsw.py	get memory statistics for the memorytracker. returns: componentstats with current memory usage.	STRING
HIGH	headroom/proxy/request_logger.py	get memory statistics for the memorytracker. returns: componentstats with current memory usage.	STRING
HIGH	headroom/proxy/semantic_cache.py	get memory statistics for the memorytracker. returns: componentstats with current memory usage.	STRING
HIGH	headroom/cache/compression_store.py	get memory statistics for the memorytracker. returns: componentstats with current memory usage.	STRING
HIGH	headroom/proxy/memory_tool_adapter.py	extract tool calls from response based on provider format.	STRING
HIGH	headroom/proxy/memory_handler.py	extract tool calls from response based on provider format.	STRING
HIGH	headroom/ccr/response_handler.py	extract tool calls from response based on provider format.	STRING
HIGH	headroom/cache/compression_feedback.py	fraction of retrievals that were full (not search).	STRING
HIGH	headroom/telemetry/toin.py	fraction of retrievals that were full (not search).	STRING
HIGH	headroom/telemetry/models.py	fraction of retrievals that were full (not search).	STRING
HIGH	headroom/integrations/strands/hooks.py	reset all tracked metrics (thread-safe). clears the metrics history and resets the total tokens saved counter. useful fo	STRING
HIGH	headroom/integrations/strands/model.py	reset all tracked metrics (thread-safe). clears the metrics history and resets the total tokens saved counter. useful fo	STRING
HIGH	headroom/integrations/agno/model.py	reset all tracked metrics (thread-safe). clears the metrics history and resets the total tokens saved counter. useful fo	STRING

Docstring Block Structure44 hits · 220 pts

Severity	File	Line	Snippet	Context
HIGH	tests/test_compression/test_evals.py	116	"""Authentication middleware for FastAPI applications. This module provides JWT-based authentication with role-based ac	STRING
HIGH	headroom/onnx_runtime.py	42	Download a file from HuggingFace Hub, preferring the local cache. Tries ``local_files_only=True`` first to avoid a	STRING
HIGH	headroom/compress.py	171	Compress messages using Headroom's full compression pipeline. This is the simplest way to use Headroom. No proxy, n	STRING
HIGH	headroom/relevance/__init__.py	76	Factory function to create a relevance scorer. Args: tier: Scorer tier to create: - "hybrid": H	STRING
HIGH	headroom/tokenizers/registry.py	134	Get tokenizer for a model. Args: model: Model name (e.g., 'gpt-4o', 'claude-3-sonnet').	STRING
HIGH	headroom/tokenizers/huggingface.py	325	Encode text to token IDs. Args: text: Text to encode. Returns: List of token I	STRING
HIGH	headroom/tokenizers/huggingface.py	344	Decode token IDs to text. Args: tokens: List of token IDs. Returns: Decoded te	STRING
HIGH	headroom/tokenizers/base.py	330	Encode text to token IDs. Optional method - not all backends support encoding. Default implementation r	STRING
HIGH	headroom/tokenizers/base.py	347	Decode token IDs to text. Optional method - not all backends support decoding. Default implementation r	STRING
HIGH	headroom/evals/prompt_comparison.py	220	Compare two prompts using an LLM judge to verify semantic equivalence. This function sends both prompts to GPT-4o (	STRING
HIGH	headroom/evals/prompt_comparison.py	594	Verify that Headroom preserves the semantic meaning of a request. This is the main entry point for verifying Headro	STRING
HIGH	headroom/learn/analyzer.py	471	Strip optional markdown fences and parse JSON. Handles raw JSON and fenced code blocks (e.g. ```json ... ```), in	STRING
HIGH	headroom/learn/analyzer.py	522	Call a locally installed CLI tool as the LLM backend. Enables keyless usage for subscription-based CLI tools that h	STRING
HIGH	headroom/memory/system.py	99	Update an existing memory with new content. Creates a new version while preserving history (supersession).	STRING
HIGH	headroom/memory/core.py	517	Supersede an existing memory with a new version. Creates a temporal chain where the old memory's validity ends	STRING
HIGH	headroom/memory/factory.py	69	Create a complete memory system from configuration. This factory function creates and initializes all memory system	STRING
HIGH	headroom/memory/factory.py	117	Create a memory store backend. Args: config: Memory system configuration. Returns: A MemorySto	STRING
HIGH	headroom/memory/factory.py	145	Create or return a cached embedder backend. The embedder is shared across every ``LocalBackend`` instance that	STRING
HIGH	headroom/memory/factory.py	231	Create a vector index backend. Args: config: Memory system configuration. Returns: A VectorInd	STRING
HIGH	headroom/memory/factory.py	324	Create a text index backend. Args: config: Memory system configuration. Returns: A TextIndex i	STRING
HIGH	headroom/memory/backends/local.py	532	Update a memory with new content (creates versioned history). Uses HierarchicalMemory.supersede() to create a n	STRING
HIGH	headroom/memory/backends/mem0_system_adapter.py	231	Update an existing memory with new content. Creates a new version while preserving history (supersession).	STRING
HIGH	headroom/memory/adapters/embedders.py	663	Call OpenAI API with retry logic for transient failures. Args: texts: List of texts to embed.	STRING
HIGH	headroom/memory/adapters/embedders.py	707	Generate an embedding for a single text. Args: text: The text to embed. Returns:	STRING
HIGH	headroom/memory/adapters/embedders.py	726	Generate embeddings for multiple texts. Automatically handles batching for large inputs. Args:	STRING
HIGH	headroom/memory/adapters/embedders.py	885	Call Ollama API with retry logic for a single text. Args: text: Text to embed. Returns:	STRING
HIGH	headroom/memory/adapters/embedders.py	940	Generate an embedding for a single text. Args: text: The text to embed. Returns:	STRING
HIGH	headroom/memory/adapters/embedders.py	959	Generate embeddings for multiple texts. Ollama API doesn't support batch embedding natively, so we make	STRING
HIGH	headroom/memory/adapters/sqlite.py	632	Supersede an existing memory with a new version. This creates a temporal chain: the old memory's valid_until is	STRING
HIGH	headroom/memory/adapters/hnsw.py	575	Search for similar memories using vector similarity. Args: filter: Vector search filter with query	STRING
HIGH	headroom/proxy/server.py	1197	Run a synchronous compression callable on the bounded executor with cancel-aware metrics. Replaces ``as	STRING
HIGH	headroom/cache/registry.py	83	Get a cache optimizer for a provider. Args: provider: Provider name (e.g., "anthropic", "o	STRING
HIGH	headroom/providers/base.py	54	Get a token counter for a specific model. Args: model: The model name. Returns:	STRING
HIGH	headroom/providers/base.py	70	Get the context window limit for a model. Args: model: The model name. Returns:	STRING
HIGH	headroom/backends/base.py	125	Send an OpenAI-format message request. Unlike send_message(), this takes OpenAI-format input and returns	STRING
HIGH	headroom/backends/base.py	148	Stream an OpenAI-format chat completion. Yields SSE-formatted strings: 'data: {...}\\n\\n' for each chunk,	STRING
HIGH	headroom/cli/_utils/parsers.py	10	Parse a duration string into a timedelta. Supported formats: - "7d" - 7 days - "2w" - 2 weeks	STRING
HIGH	headroom/transforms/spreadsheet_ingest.py	74	Load a spreadsheet file into ``{sheet_name: csv_text}``. Args: path: Path to a ``.xlsx`` or ``.xls`` file.	STRING
HIGH	headroom/transforms/code_compressor.py	106	Get a tree-sitter parser for the given language. Returns a thread-local ``tree_sitter.Parser`` instance. t	STRING
HIGH	headroom/transforms/kompress_compressor.py	1223	Compress multiple texts. Uses batched inference on GPU, sequential on CPU. On GPU (PyTorch + CUDA / MPS), runs	STRING
HIGH	headroom/compression/masks.py	88	Combine masks - preserve if EITHER mask says preserve. Useful when combining multiple structure detection strat	STRING
HIGH	headroom/compression/masks.py	111	Combine masks - preserve only if BOTH masks say preserve. Useful for being more aggressive with compression.	STRING
HIGH	headroom/rtk/installer.py	76	Download rtk binary from GitHub releases. Args: version: Version to download (e.g., "v0.42.4"). Defaults to	STRING
HIGH	headroom/pricing/registry.py	103	Estimate the cost for a given token usage. Args: model: The model name/identifier. inpu	STRING

AI Slop Vocabulary66 hits · 190 pts

Severity	File	Line	Snippet	Context
MEDIUM	codecov.yml	6	# Gate on the comprehensive unit suite (`python` flag from ci.yml's 4 test	COMMENT
MEDIUM	pyproject.toml	286	# EleutherAI lm-evaluation-harness (lm-eval), which headroom invokes as an	COMMENT
MEDIUM	pyproject.toml	291	# need the accuracy harness opt in explicitly with `pip install	COMMENT
MEDIUM	pyproject.toml	306	# ML weight a sandbox avoids; evals is a dev/test harness, not a	COMMENT
MEDIUM	pyproject.toml	476	# test-teardown thread; this is a test-harness artifact, not a production issue	COMMENT
MEDIUM	crates/headroom-parity/src/lib.rs	53	/// differing by 1 ULP). To make comparisons robust we round-trip the	COMMENT
MEDIUM	crates/headroom-py/Cargo.toml	23	# Disable the default test harness — Rust `cargo test` can't run a cdylib that	COMMENT
MEDIUM	tests/test_openai_responses_t3_replay_regression.py	173	# Simulate a harness that changed/trimmed the ancient envelope.	COMMENT
MEDIUM⚡	tests/test_request_outcome.py	146	# ── classify_client — the harness ID source ─────────────────────────	COMMENT
MEDIUM⚡	tests/test_request_outcome.py	169	# X-Client wins even when UA matches a different harness	COMMENT
MEDIUM⚡	tests/test_request_outcome.py	107	assert o.client is None # unidentified harness	CODE
MEDIUM	tests/test_request_outcome.py	214	# Bind the real method to this harness.	COMMENT
MEDIUM⚡	tests/test_google_multimodal.py	557	# Parametrized tests for comprehensive coverage	COMMENT
MEDIUM	tests/test_dashboard_cache_net_playwright.py	54	# escapes the harness to the real network.	COMMENT
LOW	tests/test_image_compressor.py	613	# URL images should just pass through since we can't extract data	COMMENT
MEDIUM	tests/test_ccr_feedback.py	304	"""get_stats returns comprehensive overview."""	STRING
MEDIUM	…test_dashboard_context_tool_availability_playwright.py	127	# escapes the harness to the real network.	COMMENT
MEDIUM⚡	tests/test_observed_wire_shapes.py	289	# HEADROOM_LOG_MESSAGES trace of the `inspect` non-litellm harness);	COMMENT
MEDIUM⚡	tests/test_observed_wire_shapes.py	332	# Section 5 — command classification is harness-agnostic (Bug A + Bug B)	COMMENT
MEDIUM⚡	tests/test_observed_wire_shapes.py	474	# The SAME read action, as each provider/harness serializes its tool call.	COMMENT
MEDIUM⚡	tests/test_observed_wire_shapes.py	511	# object — raw and harness-wrapped — is RELEASED (not protected) for compression	COMMENT
MEDIUM⚡	tests/test_observed_wire_shapes.py	534	# whatever role the harness stamps on it.	COMMENT
MEDIUM	tests/test_observed_wire_shapes.py	411	# command* so cat/sed code reads are passed verbatim on ANY model/harness.	COMMENT
MEDIUM⚡	tests/test_cross_turn_dedup.py	253	# — role:tool, role:function, or a text-harness role:user string — keyed off the	COMMENT
MEDIUM	tests/test_cross_turn_dedup.py	270	# Text-harness shape: the read output arrives as a role:user STRING after an	COMMENT
LOW	tests/test_bedrock_region.py	162	# We need to actually test the function, so let's just use the	COMMENT
MEDIUM	tests/test_proxy_hooks_regression.py	46	# from that same tokenizer so the test stays robust to any future	COMMENT
MEDIUM	tests/test_savings_ledger.py	255	# identified harness -> recorded as that client	COMMENT
MEDIUM	tests/test_savings_ledger.py	267	# unidentified harness -> falls back to "proxy"	COMMENT
MEDIUM	tests/test_dashboard_cache_ttl_playwright.py	167	# escapes the harness to the real network.	COMMENT
LOW	tests/test_evals/test_html_extraction_eval.py	308	# These should not fail (just create the evaluator)	COMMENT
MEDIUM	tests/test_evals/test_html_oss_benchmarks.py	279	# Print comprehensive results	COMMENT
MEDIUM	examples/strands_via_proxy_demo.py	9	Fix #4 Strands harness label in CLIENT_UA_MAP	STRING
MEDIUM	benchmarks/conftest.py	241	"\n\n## Guidelines\n1. Always explain your reasoning\n2. Provide code examples when helpful\n3. Consider edge ca	CODE
MEDIUM	benchmarks/comprehensive_eval.py	708	"""Run comprehensive evaluation."""	STRING
MEDIUM	e2e/init/run.py	33	# Add repo root to sys.path so the harness import works whether the file is	COMMENT
MEDIUM	e2e/init/run.py	56	# The harness restores PATH before assertions, but `headroom init` ran with a	COMMENT
MEDIUM⚡	headroom/evals/suite_runner.py	60	# Standard benchmarks via lm-eval harness (through Headroom proxy)	COMMENT
MEDIUM	headroom/evals/comprehensive_benchmark.py	63	"mmlu", # 57 subjects - comprehensive but slow	CODE
MEDIUM	headroom/evals/comprehensive_benchmark.py	467	"""Run comprehensive benchmark suite.	STRING
MEDIUM	headroom/learn/verbosity.py	273	# Backfill has_tools (a session-level property of the harness).	COMMENT
MEDIUM⚡	headroom/memory/extraction.py	39	FACT_EXTRACTION_PROMPT = """You are a comprehensive fact extractor. Your goal is to capture ALL meaningful information f	CODE
MEDIUM	headroom/proxy/server.py	3643	"""Get comprehensive proxy statistics.	STRING
MEDIUM	headroom/proxy/prometheus_metrics.py	702	# client is the harness classified from the User-Agent / X-Client	COMMENT
MEDIUM	headroom/proxy/prometheus_metrics.py	704	# when the harness is unidentified.	COMMENT
MEDIUM⚡	headroom/proxy/outcome.py	120	# client: identified harness driving the request (codex /	COMMENT
MEDIUM⚡	headroom/proxy/outcome.py	123	# User-Agent matched a known harness. Populated by handlers	COMMENT
MEDIUM⚡	headroom/proxy/outcome.py	127	# can slice by harness without a separate column. This is the	COMMENT
MEDIUM⚡	headroom/proxy/outcome.py	129	# harness visibility appears across EVERY handler with zero	COMMENT
MEDIUM	headroom/proxy/outcome.py	412	# tag-based filtering surfaces per-harness slicing for free —	COMMENT
MEDIUM	headroom/proxy/outcome.py	446	# a harness was identified — keeps the unidentified-traffic	COMMENT
MEDIUM	headroom/proxy/handlers/openai.py	4825	# Identify the WS harness before downstream auth/header rewrites.	COMMENT
MEDIUM	headroom/proxy/handlers/anthropic.py	791	# Identify the harness (codex / claude-code / aider / etc.)	COMMENT
MEDIUM	headroom/proxy/handlers/streaming.py	1065	# Identify the harness (codex / claude-code / aider / cursor /	COMMENT
MEDIUM	headroom/cache/prefix_tracker.py	321	# stripping) so it is robust to ALL per-turn transport / annotation churn —	COMMENT
LOW	headroom/backends/litellm.py	339	# Provider Registry - to add a new provider, just add an entry here!	COMMENT
MEDIUM	headroom/subscription/tracker.py	382	# PR-G2 remediation (H1): elevate to ERROR — this is config	COMMENT
MEDIUM	headroom/transforms/content_detector.py	195	# WRAPPED payload to still count as JSON: a small structural wrapper (a harness	COMMENT
MEDIUM⚡	headroom/transforms/content_router.py	3608	# keeps exact bytes to patch — regardless of how THIS harness labels	COMMENT
MEDIUM⚡	headroom/transforms/content_router.py	3610	# harnesses (Anthropic tool_use, OpenAI/Kimi `role:tool`, text-harness	COMMENT
6 more matches not shown…

Redundant / Tautological Comments116 hits · 158 pts

Severity	File	Line	Snippet	Context
LOW	tests/test_memory_system.py	26	# Check if hnswlib is available for LocalBackend tests	COMMENT
LOW	tests/test_proxy_memory_integration.py	221	# Check if response indicates tool was used	COMMENT
LOW	tests/test_proxy_memory_integration.py	281	# Check if response mentions the saved info	COMMENT
LOW⚡	tests/test_memory_integration.py	264	# Check if memory tool was called	COMMENT
LOW⚡	tests/test_memory_integration.py	271	# Check if auto-handled	COMMENT
LOW	tests/test_memory_integration.py	310	# Check if memory tool was called with pre-extraction	COMMENT
LOW	tests/test_memory_integration.py	558	# Check if search was invoked	COMMENT
LOW	tests/test_critical_gaps.py	830	# Set _last_save_time to past so elapsed > interval	COMMENT
LOW	tests/test_proxy_ccr.py	445	# Check if compression was actually performed (not skipped)	COMMENT
LOW	tests/test_hnsw_only.py	16	# Check if hnswlib is available (use lazy check to avoid SIGILL on incompatible CPUs)	COMMENT
LOW	tests/test_sqlite_vector_index.py	22	# Check if sqlite-vec is available	COMMENT
LOW	tests/test_google_multimodal_e2e.py	145	# Check if model made a function call	COMMENT
LOW	tests/test_cache/test_anthropic.py	70	# Check if cache_control was inserted	COMMENT
LOW	tests/test_evals/test_html_oss_benchmarks.py	70	# Print results	COMMENT
LOW⚡	tests/test_integrations/langchain/test_chat_model.py	1021	# Check if Ollama server is running on default port	COMMENT
LOW	tests/test_integrations/langchain/test_chat_model.py	18	# Check if LangChain is available	COMMENT
LOW	tests/test_integrations/langchain/test_extended.py	17	# Check if LangChain is available	COMMENT
LOW	tests/test_integrations/langchain/test_streaming.py	16	# Check if LangChain is available	COMMENT
LOW	tests/test_integrations/langchain/test_langgraph.py	15	# Check if LangChain is available	COMMENT
LOW	tests/test_integrations/langchain/test_memory.py	15	# Check if LangChain is available	COMMENT
LOW	tests/test_integrations/langchain/test_retriever.py	15	# Check if LangChain is available	COMMENT
LOW	tests/test_integrations/langchain/test_agents.py	16	# Check if LangChain is available	COMMENT
LOW	tests/test_integrations/agno/test_model.py	15	# Check if Agno is available	COMMENT
LOW	tests/test_integrations/agno/test_model.py	1032	# Check if Ollama server is running on default port	COMMENT
LOW	tests/test_integrations/agno/test_hooks.py	14	# Check if Agno is available	COMMENT
LOW⚡	tests/test_memory/test_factory.py	29	# Check if hnswlib is available (most factory tests require it)	COMMENT
LOW⚡	tests/test_memory/test_hierarchical.py	659	# Check if hnswlib is available (use lazy check to avoid SIGILL on incompatible CPUs)	COMMENT
LOW	tests/test_memory/test_easy.py	29	# Check if hnswlib is available (local backend requires it)	COMMENT
LOW	tests/test_memory/test_core_operations.py	36	# Check if hnswlib is available (HierarchicalMemory requires it)	COMMENT
LOW	tests/test_cli/test_mcp.py	24	# Check if MCP SDK is available	COMMENT
LOW	tests/integrations/test_strands/test_model.py	23	# Check if strands-agents is installed	COMMENT
LOW	tests/integrations/test_strands/test_model_unit.py	16	# Check if strands-agents is installed for proper skip handling	COMMENT
LOW	tests/integrations/test_strands/test_hooks.py	23	# Check if strands-agents is installed	COMMENT
LOW	tests/integrations/test_strands/test_hooks_unit.py	18	# Check if strands-agents is installed for proper skip handling	COMMENT
LOW	examples/strands_bedrock_demo.py	646	# Display results	COMMENT
LOW	examples/strands_bedrock_demo.py	816	# Display results	COMMENT
LOW	examples/langchain_demo/run_comparison.py	195	# Check if done	COMMENT
LOW	examples/langchain_demo/run_comparison.py	299	# Check if done	COMMENT
LOW⚡	examples/deployment/macos-launchagent/uninstall.sh	67	# Check if we're on macOS	COMMENT
LOW⚡	examples/deployment/macos-launchagent/uninstall.sh	74	# Check if LaunchAgent is installed	COMMENT
LOW⚡	examples/deployment/macos-launchagent/uninstall.sh	81	# Check if service is running and stop it	COMMENT
LOW	examples/deployment/macos-launchagent/install.sh	75	# Check if we're on macOS	COMMENT
LOW	examples/deployment/macos-launchagent/install.sh	81	# Check if headroom is installed	COMMENT
LOW	examples/deployment/macos-launchagent/install.sh	96	# Check if service is already installed	COMMENT
LOW	examples/deployment/macos-launchagent/install.sh	100	# Check if service is running	COMMENT
LOW	examples/deployment/macos-launchagent/install.sh	130	# Check if port is in use	COMMENT
LOW	examples/deployment/macos-launchagent/install.sh	189	# Check if port is listening	COMMENT
LOW	…ples/deployment/macos-launchagent/shell-integration.sh	32	# Check if proxy is running (fast path using lsof)	COMMENT
LOW	…ples/deployment/macos-launchagent/shell-integration.sh	41	# Check if LaunchAgent is installed	COMMENT
LOW⚡	benchmarks/adversarial_ccr_tests.py	658	# Check if error items with unicode were preserved	COMMENT
LOW⚡	benchmarks/adversarial_ccr_tests.py	1505	# Check if we kept the TARGET item	COMMENT
LOW⚡	benchmarks/adversarial_ccr_tests.py	1614	# Check if null-byte items were preserved (they have status=error)	COMMENT
LOW	benchmarks/adversarial_ccr_tests.py	332	# Check if anomaly was preserved	COMMENT
LOW	benchmarks/adversarial_ccr_tests.py	618	# Check if our compression worked despite fake markers	COMMENT
LOW	benchmarks/adversarial_ccr_tests.py	1001	# Check if the nested error was preserved	COMMENT
LOW	benchmarks/adversarial_ccr_tests.py	1121	# Check if future timestamp was preserved	COMMENT
LOW	benchmarks/compression_benchmark.py	537	# Check if most key terms appear in answer	COMMENT
LOW	benchmarks/run_benchmarks.py	260	# Write file	COMMENT
LOW	benchmarks/ccr_regression_benchmark.py	150	# Check if ALL errors were retained	COMMENT
LOW	benchmarks/ccr_regression_benchmark.py	225	# Check if target UUID was found	COMMENT
56 more matches not shown…

Modern AI Meta-Vocabulary46 hits · 145 pts

Severity	File	Line	Snippet	Context
MEDIUM	CHANGELOG.md	115	* wrap/doctor: make the Claude Remote Control gate warning accurate and stop it firing for users who never had the f	COMMENT
MEDIUM	crates/headroom-core/src/ccr/mod.rs	63	/// (`CCRConfig.store_ttl_seconds`). Session-scale: agentic sessions	COMMENT
MEDIUM	REALIGNMENT/04-phase-B-live-zone.md	88	- MessageScorer Rust port retirement: PR #338 and #343 (April 2026) ported MessageScorer to Rust. That work becomes	CODE
MEDIUM	REALIGNMENT/12-decisions-needed.md	35	The PR #338 / #343 port (April 2026) was investment in the wrong abstraction (per Agent G's audit: scoring's only consum	CODE
MEDIUM⚡	tests/test_proxy_warmup.py	102	# Startup orchestration tests — use HeadroomProxy + stubbed transforms	COMMENT
MEDIUM	tests/test_openai_codex_ws_lifecycle.py	768	# out is a cancel from the handler's orchestration.	COMMENT
MEDIUM	tests/test_memory_handler_native_ops.py	848	# handles its own context window. See ``MemoryQuery``.	COMMENT
MEDIUM⚡	tests/test_relevance_split.py	202	# the embedding model on the request thread (the swap happens later, in the	COMMENT
MEDIUM⚡	tests/test_memory/test_traffic_learner.py	2105	# Malformed input — we'd rather drop than persist scaffolding.	COMMENT
MEDIUM⚡	tests/test_cli/test_wrap_helpers.py	840	# _resolve_1m_model — 1M context window suffix logic (#1158).	COMMENT
MEDIUM	wiki/langchain.md	396	### Example 2: RAG Pipeline with Document Filtering	COMMENT
MEDIUM⚡	benchmarks/agent_cost_benchmark.py	348	# SCENARIO 3: RAG Context Scaling	COMMENT
MEDIUM	benchmarks/agent_cost_benchmark.py	362	# Generate RAG conversation with ~50K tokens of context	COMMENT
MEDIUM⚡	benchmarks/bench_latency.py	114	content_type: str # json, code, text, logs, agentic, rag	CODE
MEDIUM	benchmarks/bench_latency.py	589	# Set a model_limit large enough to exercise compression on big agentic contexts	COMMENT
MEDIUM	benchmarks/bench_latency.py	601	# --- RAG conversations ---	COMMENT
MEDIUM	benchmarks/bench_latency.py	864	# --- Per-Transform Breakdown (for agentic/rag scenarios) ---	COMMENT
MEDIUM	.github/workflows/ci.yml	7	# prefetch-model — download the embedding model ONCE (authenticated), warm shared cache	COMMENT
MEDIUM	headroom/config.py	540	# Session-scale TTL. The original 5-minute default predates agentic	COMMENT
MEDIUM	headroom/__init__.py	189	# Shared context for multi-agent workflows	COMMENT
MEDIUM	headroom/parser.py	39	# Patterns for RAG detection (best effort)	COMMENT
MEDIUM	headroom/parser.py	260	# Check if this looks like RAG content	COMMENT
MEDIUM⚡	headroom/evals/datasets.py	49	# RAG / RETRIEVAL DATASETS	COMMENT
MEDIUM⚡	headroom/evals/datasets.py	1148	# RAG/Retrieval	COMMENT
MEDIUM	headroom/evals/README.md	176	### RAG / Retrieval	COMMENT
MEDIUM	headroom/evals/__main__.py	389	python -m headroom.evals benchmark --dataset rag # Run all RAG datasets	CODE
MEDIUM	headroom/learn/models.py	162	# guards against a detected repeated pattern. Loop guardrails are ranked	COMMENT
MEDIUM⚡	headroom/learn/analyzer.py	200	# Weight loop guardrails above one-off rules using MEASURED waste.	COMMENT
MEDIUM⚡	headroom/memory/traffic_learner.py	958	# implementation matched scaffolding ("don't mention this	COMMENT
MEDIUM	headroom/proxy/server.py	775	# code/RAG context — see issue #454).	COMMENT
MEDIUM	headroom/proxy/server.py	2848	# remain the unauthenticated probes for orchestration health.	COMMENT
MEDIUM	headroom/proxy/models.py	219	# of input lives in user messages (pasted code/text, RAG context) and the	COMMENT
MEDIUM⚡	headroom/proxy/helpers.py	2612	# local context window — tens of K tokens. That decision is made client-side	COMMENT
MEDIUM⚡	headroom/proxy/helpers.py	2724	# Anthropic then excludes deferred tools from the context window (they stop	COMMENT
MEDIUM	headroom/proxy/handlers/anthropic.py	2256	# them from the context window — they stop counting as input tokens until	COMMENT
MEDIUM	headroom/providers/openai.py	72	# OpenAI context window limits	COMMENT
MEDIUM	headroom/providers/anthropic.py	96	# Claude Sonnet 4.6 - 1M context window	COMMENT
MEDIUM	headroom/providers/claude/runtime.py	12	# every one into its context window) when ANTHROPIC_BASE_URL is a custom host	COMMENT
MEDIUM	headroom/providers/claude/runtime.py	30	# family as #746 (on-demand tool loading) and #1158 (1M context window), both of	COMMENT
MEDIUM	headroom/providers/claude/install.py	20	# its context window (GH #746) — breaking sub-agents and forcing compaction.	COMMENT
MEDIUM	headroom/cli/init.py	170	# all into its context window — overflowing it (breaks sub-agent spawns,	COMMENT
MEDIUM	headroom/cli/wrap.py	169	# inflates the local context window by tens of K tokens. Setting the env var	COMMENT
MEDIUM	headroom/cli/wrap.py	179	# 1M context window for `wrap claude` (#1158). Claude Code only sends the	COMMENT
MEDIUM	headroom/cli/wrap.py	4208	# Issue #1158: opt-in 1M context window. Claude Code only sends the	COMMENT
MEDIUM	headroom/cli/wrap.py	3932	headroom wrap claude --1m # Preserve the 1M context window	STRING
MEDIUM	headroom/transforms/content_router.py	1144	# embedding model is pre-warmed in the background (BM25 scores until it's	COMMENT

Self-Referential Comments43 hits · 134 pts

Severity	File	Line	Snippet	Context
MEDIUM	crates/headroom-core/config/pipeline.toml	3	# This file is embedded into the headroom-core binary via `include_str!`	COMMENT
MEDIUM	tests/test_security_validations.py	109	# Create a memory with safe metadata	COMMENT
MEDIUM⚡	tests/test_memory_system.py	642	# Create a simple graph: e1 -> e2 -> e3	COMMENT
MEDIUM⚡	tests/test_memory_system.py	802	# Create a chain longer than max_depth	COMMENT
MEDIUM	tests/test_memory_system.py	665	# Create a chain: e1 -> e2 -> e3 -> e4	COMMENT
MEDIUM	tests/test_memory_system.py	1688	# Create a graph structure	COMMENT
MEDIUM⚡	tests/test_image_compression.py	256	# Create a simple test image	COMMENT
MEDIUM⚡	tests/test_image_compressor.py	61	# Create a simple 100x100 red image	COMMENT
MEDIUM	tests/test_image_compressor.py	74	# Create a 1024x1024 image with some pattern	COMMENT
MEDIUM	tests/test_critical_fixes.py	110	# Create a signature using the correct factory method	COMMENT
MEDIUM	tests/test_critical_gaps.py	69	# Create a pattern	COMMENT
MEDIUM	tests/test_toin_feedback.py	338	# Create a minimal proxy to test the method	COMMENT
MEDIUM	tests/test_proxy_batch_integration.py	76	# Create a list of items that will be compressible	COMMENT
MEDIUM⚡	tests/test_backend_bugs.py	617	# Create an async iterator for the mock streaming response	COMMENT
MEDIUM	tests/test_acceptance.py	16	# Create a shared provider for tests	COMMENT
MEDIUM	tests/repro_unsendable_panic.py	41	# Create the parser on the main thread	COMMENT
MEDIUM	tests/test_memory_tracker_integration.py	414	# Create a component that reports large size	COMMENT
MEDIUM	tests/test_memory_sync.py	232	# Create an existing MEMORY.md	COMMENT
MEDIUM⚡	tests/test_adapter_hooks.py	213	# Create a new TOIN instance with same backend — should load patterns	COMMENT
MEDIUM	tests/test_cli_update.py	247	# Create a fake .pyd file	COMMENT
MEDIUM	tests/test_proxy_streaming_resilience.py	247	# Create a realistic HTTP 429 error	COMMENT
MEDIUM	tests/test_proxy_compress_endpoint.py	175	# Create a large repetitive tool output that should be compressible	COMMENT
MEDIUM	tests/test_cache/test_client_integration.py	294	# Create a spy optimizer to track calls	COMMENT
MEDIUM	tests/test_cache/test_client_integration.py	618	# Create a conversation that will trigger some optimization	COMMENT
MEDIUM	tests/test_integrations/langchain/test_chat_model.py	1213	# Create a conversation with repetitive content that should be compressed	COMMENT
MEDIUM⚡	tests/test_integrations/agno/test_model.py	877	# Create a real OpenAIChat model (doesn't need API key for instantiation)	COMMENT
MEDIUM	tests/test_integrations/agno/test_model.py	55	# Create a proper ModelResponse that Agno's response() can process	COMMENT
MEDIUM	tests/test_memory/test_hierarchical.py	344	# Create a chain: v1 -> v2 -> v3	COMMENT
MEDIUM	tests/test_memory/test_hierarchical.py	594	# Create a supersession chain	COMMENT
MEDIUM	tests/test_memory/test_budget.py	124	# Create a manager with git check disabled but file-exists check active	COMMENT
MEDIUM	tests/test_memory/test_budget.py	147	# Create the referenced file	COMMENT
MEDIUM	tests/test_transforms/test_diff_compressor.py	124	# Create a diff with many hunks	COMMENT
MEDIUM	e2e/_lib/make_shim.sh	2	# Create a noop executable shim at $2/$1 suitable for use in PATH during	COMMENT
MEDIUM	headroom/evals/batch_compression_eval.py	737	# Create a long document	COMMENT
MEDIUM	headroom/evals/runners/compression_only.py	294	# Create a document with specific facts embedded	COMMENT
MEDIUM	headroom/memory/backends/direct_mem0.py	587	# Create the background task	STRING
MEDIUM	headroom/memory/adapters/fts5.py	412	# Create a minimal Memory object from FTS data	COMMENT
MEDIUM	headroom/cache/compression_store.py	765	# Create a synthetic "success" event that we'll process later	COMMENT
MEDIUM	headroom/backends/anyllm.py	49	# Create the AnyLLM instance once and reuse. api_key/api_base are only	COMMENT
MEDIUM	headroom/integrations/asgi.py	164	# Create a new receive that returns the (possibly modified) body	COMMENT
MEDIUM	headroom/integrations/strands/hooks.py	10	# Create the hook provider	STRING
MEDIUM	headroom/transforms/content_router.py	639	# Create a deterministic structure hash based on content type	COMMENT
MEDIUM	headroom/transforms/content_router.py	1444	# Create a content-type signature	COMMENT

Structural Annotation Overuse61 hits · 111 pts

Severity	File	Line	Snippet	Context
LOW⚡	crates/headroom-core/tests/ccr_backends.rs	136	// Step 1: same payload yields byte-equal keys.	COMMENT
LOW⚡	crates/headroom-core/tests/ccr_backends.rs	139	// Step 2: store in sqlite, mirror to in-memory under the same	COMMENT
LOW⚡	tests/test_proxy_count_tokens_integration.py	464	# Step 1: Count tokens with large content	STRING
LOW⚡	tests/test_proxy_count_tokens_integration.py	470	# Step 2: Make countTokens request	STRING
LOW⚡	tests/test_proxy_count_tokens_integration.py	479	# Step 3: Verify stats updated	STRING
LOW⚡	tests/test_proxy_count_tokens_integration.py	483	# Step 4: Verify tokens saved is tracked (may be negative for small overhead)	STRING
LOW⚡	tests/test_proxy_batch_integration.py	150	# Step 1: Create JSONL content with compressible messages	COMMENT
LOW	tests/test_proxy_batch_integration.py	169	# Step 2: Upload the JSONL file directly to OpenAI	COMMENT
LOW	tests/test_proxy_batch_integration.py	183	# Step 3: Create batch through proxy with compression	COMMENT
LOW	tests/test_proxy_batch_integration.py	221	# Step 4: Cancel the batch to avoid costs	COMMENT
LOW⚡	tests/test_proxy_ccr.py	599	# Step 2: Retrieve through proxy endpoint (by hash → full content)	COMMENT
LOW⚡	tests/test_proxy_ccr.py	609	# Step 3: Verify TOIN learned	COMMENT
LOW⚡	tests/test_proxy_ccr.py	620	# Step 4: Verify field semantics were learned	COMMENT
LOW⚡	tests/test_proxy_ccr.py	627	# Step 5: PR-B5 retired the request-time recommendation API in favor of	COMMENT
LOW	wiki/ARCHITECTURE.md	304	### Step 1: You call the API	COMMENT
LOW	wiki/ARCHITECTURE.md	319	### Step 2: HeadroomClient intercepts	COMMENT
LOW	wiki/ARCHITECTURE.md	331	### Step 3: Transform Pipeline runs	COMMENT
LOW	wiki/ARCHITECTURE.md	362	### Step 4: Call real API	COMMENT
LOW	wiki/ARCHITECTURE.md	371	### Step 5: Log metrics and return	COMMENT
LOW⚡	wiki/quickstart.md	50	### Step 1: Start the Proxy	COMMENT
LOW⚡	wiki/quickstart.md	56	### Step 2: Verify It's Running	COMMENT
LOW⚡	wiki/quickstart.md	63	### Step 3: Point Your Client	COMMENT
LOW	wiki/quickstart.md	80	### Step 4: Check Savings	COMMENT
LOW	wiki/macos-deployment.md	106	### Step 1: Create Log Directory	COMMENT
LOW	wiki/macos-deployment.md	112	### Step 2: Generate LaunchAgent Plist	COMMENT
LOW	wiki/macos-deployment.md	139	### Step 3: Load the LaunchAgent	COMMENT
LOW	wiki/macos-deployment.md	145	### Step 4: Verify Service	COMMENT
LOW⚡	benchmarks/ccr_regression_benchmark.py	642	# Step 1: Compress	COMMENT
LOW⚡	benchmarks/ccr_regression_benchmark.py	647	# Step 2: Cache	COMMENT
LOW⚡	benchmarks/ccr_regression_benchmark.py	656	# Step 3: Record compression in feedback	COMMENT
LOW⚡	benchmarks/ccr_regression_benchmark.py	659	# Step 4: Retrieve and search	COMMENT
LOW⚡	benchmarks/ccr_regression_benchmark.py	663	# Step 5: Process feedback	COMMENT
LOW⚡	benchmarks/scenarios/conversations.py	502	### Step 1: Setup	COMMENT
LOW⚡	benchmarks/scenarios/conversations.py	508	### Step 2: Implementation	COMMENT
LOW	benchmarks/scenarios/conversations.py	523	### Step 3: Testing	COMMENT
LOW⚡	headroom/memory/budget.py	107	# Step 1: Apply temporal decay to importance scores	COMMENT
LOW⚡	headroom/memory/budget.py	110	# Step 2: Detect and flag stale memories	COMMENT
LOW⚡	headroom/memory/budget.py	114	# Step 3: Merge similar memories	COMMENT
LOW⚡	headroom/memory/budget.py	118	# Step 4: Rank by score and apply budget	COMMENT
LOW⚡	headroom/proxy/handlers/batch.py	893	# Step 3: Create compressed JSONL content	COMMENT
LOW⚡	headroom/proxy/handlers/batch.py	896	# Step 4: Upload compressed file to OpenAI	COMMENT
LOW	headroom/proxy/handlers/batch.py	861	# Step 1: Download the input file from OpenAI	COMMENT
LOW	headroom/proxy/handlers/batch.py	877	# Step 2: Parse and compress each line	COMMENT
LOW	headroom/proxy/handlers/batch.py	914	# Step 5: Create batch with compressed file	COMMENT
LOW⚡	headroom/cache/anthropic.py	135	# Step 1: Analyze content sections	COMMENT
LOW⚡	headroom/cache/anthropic.py	138	# Step 2: Stabilize prefix	COMMENT
LOW⚡	headroom/cache/anthropic.py	144	# Step 3: Plan breakpoint placement	COMMENT
LOW⚡	headroom/cache/anthropic.py	148	# Step 4: Insert cache_control blocks	COMMENT
LOW⚡	headroom/cache/anthropic.py	153	# Step 5: Compute metrics	COMMENT
LOW⚡	headroom/image/trained_router.py	346	# Step 1: Classify query with trained model	COMMENT
LOW⚡	headroom/image/trained_router.py	349	# Step 2: Analyze image with SigLIP (if enabled)	COMMENT
LOW⚡	headroom/image/trained_router.py	355	# Step 3: Combine signals for final decision	COMMENT
LOW⚡	headroom/image/tile_optimizer.py	60	# Step 1: scale so max dimension ≤ 2048	COMMENT
LOW⚡	headroom/image/tile_optimizer.py	67	# Step 2: scale so shortest side ≤ 768	COMMENT
LOW⚡	headroom/image/tile_optimizer.py	74	# Step 3: count 512×512 tiles	COMMENT
LOW⚡	headroom/image/compressor.py	649	# Step 2: ML-based technique routing	COMMENT
LOW⚡	headroom/image/compressor.py	699	# Step 3: Apply compression technique	COMMENT
LOW	headroom/image/compressor.py	634	# Step 1: Tile-boundary optimization (always safe, pure math)	COMMENT
LOW⚡	headroom/telemetry/toin.py	1438	# Step 1: Serialize under lock (fast in-memory operation)	COMMENT
LOW⚡	headroom/telemetry/toin.py	1442	# Step 2: Write outside lock (slow I/O operation)	COMMENT
1 more matches not shown…

Hallucination Indicators9 hits · 95 pts

Severity	File	Line	Snippet	Context
CRITICAL	tests/test_proxy_savings_history.py	1162	client.app.state.proxy.metrics.savings_tracker.flush()	CODE
CRITICAL	tests/test_openai_chat_turn_hooks.py	207	logs = client.app.state.proxy.logger.get_recent(10)	CODE
CRITICAL	tests/test_compression/test_evals.py	357	at com.example.app.client.InventoryClient.updateStock(InventoryClient.java:87)	STRING
CRITICAL	tests/test_compression/test_evals.py	358	at com.example.app.service.InventoryService.decrementStock(InventoryService.java:156)	STRING
CRITICAL	benchmarks/headroom_worst_case_benchmark.py	290	at org.springframework.web.servlet.FrameworkServlet.service(FrameworkServlet.java:897)	CODE
CRITICAL⚡	benchmarks/comprehensive_eval.py	139	at org.springframework.jdbc.datasource.DataSourceUtils.getConnection(DataSourceUtils.java:82)	CODE
CRITICAL⚡	benchmarks/comprehensive_eval.py	140	at org.springframework.jdbc.core.JdbcTemplate.execute(JdbcTemplate.java:376)	CODE
CRITICAL⚡	benchmarks/comprehensive_eval.py	143	at org.apache.commons.dbcp2.BasicDataSource.getConnection(BasicDataSource.java:1421)	CODE
CRITICAL	headroom/memory/wrapper.py	243	response = self._wrapper._client.chat.completions.create(**kwargs)	CODE

Verbosity Indicators48 hits · 90 pts

Severity	File	Line	Snippet	Context
LOW⚡	crates/headroom-core/tests/ccr_backends.rs	136	// Step 1: same payload yields byte-equal keys.	COMMENT
LOW⚡	crates/headroom-core/tests/ccr_backends.rs	139	// Step 2: store in sqlite, mirror to in-memory under the same	COMMENT
LOW⚡	tests/test_proxy_count_tokens_integration.py	464	# Step 1: Count tokens with large content	STRING
LOW⚡	tests/test_proxy_count_tokens_integration.py	470	# Step 2: Make countTokens request	STRING
LOW⚡	tests/test_proxy_count_tokens_integration.py	479	# Step 3: Verify stats updated	STRING
LOW⚡	tests/test_proxy_count_tokens_integration.py	483	# Step 4: Verify tokens saved is tracked (may be negative for small overhead)	STRING
LOW⚡	tests/test_proxy_batch_integration.py	150	# Step 1: Create JSONL content with compressible messages	COMMENT
LOW	tests/test_proxy_batch_integration.py	169	# Step 2: Upload the JSONL file directly to OpenAI	COMMENT
LOW	tests/test_proxy_batch_integration.py	183	# Step 3: Create batch through proxy with compression	COMMENT
LOW	tests/test_proxy_batch_integration.py	221	# Step 4: Cancel the batch to avoid costs	COMMENT
LOW⚡	tests/test_proxy_ccr.py	599	# Step 2: Retrieve through proxy endpoint (by hash → full content)	COMMENT
LOW⚡	tests/test_proxy_ccr.py	609	# Step 3: Verify TOIN learned	COMMENT
LOW⚡	tests/test_proxy_ccr.py	620	# Step 4: Verify field semantics were learned	COMMENT
LOW⚡	tests/test_proxy_ccr.py	627	# Step 5: PR-B5 retired the request-time recommendation API in favor of	COMMENT
LOW⚡	benchmarks/ccr_regression_benchmark.py	642	# Step 1: Compress	COMMENT
LOW⚡	benchmarks/ccr_regression_benchmark.py	647	# Step 2: Cache	COMMENT
LOW⚡	benchmarks/ccr_regression_benchmark.py	656	# Step 3: Record compression in feedback	COMMENT
LOW⚡	benchmarks/ccr_regression_benchmark.py	659	# Step 4: Retrieve and search	COMMENT
LOW⚡	benchmarks/ccr_regression_benchmark.py	663	# Step 5: Process feedback	COMMENT
LOW⚡	benchmarks/scenarios/conversations.py	502	### Step 1: Setup	COMMENT
LOW⚡	benchmarks/scenarios/conversations.py	508	### Step 2: Implementation	COMMENT
LOW	benchmarks/scenarios/conversations.py	523	### Step 3: Testing	COMMENT
LOW⚡	headroom/memory/budget.py	107	# Step 1: Apply temporal decay to importance scores	COMMENT
LOW⚡	headroom/memory/budget.py	110	# Step 2: Detect and flag stale memories	COMMENT
LOW⚡	headroom/memory/budget.py	114	# Step 3: Merge similar memories	COMMENT
LOW⚡	headroom/memory/budget.py	118	# Step 4: Rank by score and apply budget	COMMENT
LOW⚡	headroom/proxy/handlers/batch.py	893	# Step 3: Create compressed JSONL content	COMMENT
LOW⚡	headroom/proxy/handlers/batch.py	896	# Step 4: Upload compressed file to OpenAI	COMMENT
LOW	headroom/proxy/handlers/batch.py	861	# Step 1: Download the input file from OpenAI	COMMENT
LOW	headroom/proxy/handlers/batch.py	877	# Step 2: Parse and compress each line	COMMENT
LOW	headroom/proxy/handlers/batch.py	914	# Step 5: Create batch with compressed file	COMMENT
LOW⚡	headroom/cache/anthropic.py	135	# Step 1: Analyze content sections	COMMENT
LOW⚡	headroom/cache/anthropic.py	138	# Step 2: Stabilize prefix	COMMENT
LOW⚡	headroom/cache/anthropic.py	144	# Step 3: Plan breakpoint placement	COMMENT
LOW⚡	headroom/cache/anthropic.py	148	# Step 4: Insert cache_control blocks	COMMENT
LOW⚡	headroom/cache/anthropic.py	153	# Step 5: Compute metrics	COMMENT
LOW⚡	headroom/image/trained_router.py	346	# Step 1: Classify query with trained model	COMMENT
LOW⚡	headroom/image/trained_router.py	349	# Step 2: Analyze image with SigLIP (if enabled)	COMMENT
LOW⚡	headroom/image/trained_router.py	355	# Step 3: Combine signals for final decision	COMMENT
LOW⚡	headroom/image/tile_optimizer.py	60	# Step 1: scale so max dimension ≤ 2048	COMMENT
LOW⚡	headroom/image/tile_optimizer.py	67	# Step 2: scale so shortest side ≤ 768	COMMENT
LOW⚡	headroom/image/tile_optimizer.py	74	# Step 3: count 512×512 tiles	COMMENT
LOW⚡	headroom/image/compressor.py	649	# Step 2: ML-based technique routing	COMMENT
LOW⚡	headroom/image/compressor.py	699	# Step 3: Apply compression technique	COMMENT
LOW	headroom/image/compressor.py	634	# Step 1: Tile-boundary optimization (always safe, pure math)	COMMENT
LOW⚡	headroom/telemetry/toin.py	1438	# Step 1: Serialize under lock (fast in-memory operation)	COMMENT
LOW⚡	headroom/telemetry/toin.py	1442	# Step 2: Write outside lock (slow I/O operation)	COMMENT
LOW⚡	headroom/telemetry/toin.py	1446	# Step 3: Update state under lock (fast)	COMMENT

Magic Placeholder Names10 hits · 48 pts

Severity	File	Line	Snippet	Context
HIGH	tests/test_transforms/test_html_extractor.py	466	<pre><code>Authorization: Bearer YOUR_API_KEY</code></pre>	CODE
HIGH	wiki/typescript-sdk.md	59	apiKey: 'your-api-key', // optional, for authenticated endpoints	CODE
HIGH	wiki/typescript-sdk.md	92	apiKey: 'your-api-key',	CODE
HIGH	wiki/configuration.md	431	export HEADROOM_API_KEY=your-api-key	CODE
HIGH	wiki/configuration.md	443	apiKey: 'your-api-key',	CODE
HIGH	wiki/proxy.md	317	api_key="your-api-key", # Still needed for upstream	CODE
HIGH	sdk/typescript/README.md	363	apiKey: 'your-api-key', // optional, for authenticated endpoints	CODE
HIGH	benchmarks/scenarios/conversations.py	542	Authorization: Bearer <your-api-key>	CODE
HIGH	headroom/evals/html_extraction.py	557	<pre><code>Authorization: Bearer YOUR_API_KEY</code></pre>	CODE
HIGH	headroom/providers/google.py	14	genai.configure(api_key="your-api-key")	STRING

Fake / Example Data36 hits · 32 pts

Severity	File	Line	Snippet	Context
LOW	crates/headroom-core/tests/live_zone_ccr.rs	27	"notes": "lorem ipsum dolor sit amet, consectetur adipiscing elit",	CODE
LOW	crates/headroom-core/tests/live_zone_ccr.rs	27	"notes": "lorem ipsum dolor sit amet, consectetur adipiscing elit",	CODE
LOW	…/headroom-core/src/transforms/smart_crusher/anchors.rs	295	assert!(anchors.contains("user@example.com"));	CODE
LOW	…-proxy/tests/integration_e3_anthropic_cache_control.rs	84	.header("x-api-key", "sk-ant-api01-fake-key")	CODE
LOW	…-proxy/tests/integration_e3_anthropic_cache_control.rs	174	.header("x-api-key", "sk-ant-api01-fake-key")	CODE
LOW	…-proxy/tests/integration_e3_anthropic_cache_control.rs	278	.header("x-api-key", "sk-ant-api01-fake-key")	CODE
LOW	…-proxy/tests/integration_e3_anthropic_cache_control.rs	374	.header("x-api-key", "sk-ant-api01-fake-key")	CODE
LOW⚡	tests/test_memory_system.py	1424	entities=["Alice", "Acme Corp"],	CODE
LOW⚡	tests/test_memory_system.py	1428	assert "Acme Corp" in memory.entity_refs	CODE
LOW⚡	tests/test_memory_system.py	1433	acme_entity = await graph.get_entity_by_name("alice", "Acme Corp")	CODE
LOW⚡	tests/test_memory_system.py	1445	entities=["Alice", "Acme Corp"],	CODE
LOW⚡	tests/test_memory_system.py	1446	relationships=[{"source": "Alice", "target": "Acme Corp", "type": "works_at"}],	CODE
LOW⚡	tests/test_memory_integration.py	412	entities=["Alice", "Acme Corp"],	CODE
LOW⚡	tests/test_memory_integration.py	415	{"entity": "Acme Corp", "entity_type": "organization"},	CODE
LOW⚡	tests/test_memory_integration.py	421	"destination": "Acme Corp",	CODE
LOW⚡	tests/test_memory_integration.py	430	entities=["Acme Corp", "San Francisco"],	CODE
LOW⚡	tests/test_memory_integration.py	432	{"entity": "Acme Corp", "entity_type": "organization"},	CODE
LOW	tests/test_quality_retention.py	325	"snippet": "Lorem ipsum " * 20,	CODE
LOW⚡	tests/test_backend_bugs.py	778	"x-api-key": "sk-ant-dummy-key",	CODE
LOW⚡	tests/test_backend_bugs.py	779	"authorization": "Bearer sk-ant-dummy-key",	CODE
LOW	tests/parity/recorder.py	571	"Lorem ipsum dolor sit amet, consectetur adipiscing elit.",	STRING
LOW	tests/parity/recorder.py	571	"Lorem ipsum dolor sit amet, consectetur adipiscing elit.",	STRING
LOW	tests/test_transforms/test_html_extractor.py	218	<meta name="author" content="John Doe">	CODE
LOW	tests/test_compression/test_llm_eval.py	201	email="admin@example.com",	CODE
LOW	wiki/langchain.md	525	("placeholder", "{agent_scratchpad}"),	CODE
LOW	examples/mcp_demo/run_agent_eval.py	180	"email": "test@test.com",	CODE
LOW	examples/mcp_demo/mock_mcp_servers.py	179	"body": "Lorem ipsum dolor sit amet..." if random.random() > 0.5 else "",	CODE
LOW	examples/mcp_demo/mock_mcp_servers.py	179	"body": "Lorem ipsum dolor sit amet..." if random.random() > 0.5 else "",	CODE
LOW⚡	scripts/replay_codex_ws_load.py	126	"Lorem ipsum dolor sit amet, consectetur adipiscing elit. "	CODE
LOW⚡	scripts/replay_codex_ws_load.py	126	"Lorem ipsum dolor sit amet, consectetur adipiscing elit. "	CODE
LOW	scripts/replay_codex_ws_load.py	154	# plain_text_like / unknown / jsonl_like → lorem ipsum is fine as a	COMMENT
LOW	headroom/evals/datasets.py	1040	"name": "Acme Corp",	CODE
LOW	headroom/memory/tools.py	326	Example: [{"source": "user", "relationship": "works_at", "destination": "Acme Corp"}]	CODE
LOW⚡	headroom/memory/extraction.py	106	{"entity": "Acme Corp", "entity_type": "organization"},	CODE
LOW	headroom/memory/backends/local.py	98	entities=["Alice", "Acme Corp"],	STRING
LOW	headroom/memory/backends/local.py	99	relationships=[{"source": "Alice", "target": "Acme Corp", "type": "works_at"}],	STRING

AI Structural Patterns28 hits · 22 pts

Severity	File	Line	Context
LOW	plugins/headroom-oauth2/src/headroom_oauth2/provider.py	41	CODE
LOW	tests/test_memory_system.py	898	CODE
LOW	tests/test_proxy_compression_executor.py	114	CODE
LOW	headroom/pipeline.py	132	CODE
LOW	headroom/evals/prompt_comparison.py	537	CODE
LOW	headroom/memory/core.py	131	CODE
LOW	headroom/memory/easy.py	108	CODE
LOW	headroom/memory/backends/direct_mem0.py	531	CODE
LOW	headroom/memory/backends/local.py	231	CODE
LOW	headroom/memory/adapters/embedders.py	155	CODE
LOW	headroom/memory/adapters/hnsw.py	223	CODE
LOW	headroom/proxy/savings_tracker.py	615	CODE
LOW	headroom/proxy/server.py	1769	CODE
LOW	headroom/proxy/prometheus_metrics.py	570	CODE
LOW	headroom/proxy/outcome.py	190	CODE
LOW	headroom/proxy/helpers.py	167	CODE
LOW	headroom/proxy/handlers/streaming.py	787	CODE
LOW	headroom/proxy/handlers/streaming.py	958	CODE
LOW	headroom/cache/semantic.py	412	CODE
LOW	headroom/cache/compression_store.py	253	CODE
LOW	headroom/prediction/feature_extractor.py	2298	CODE
LOW	headroom/cli/wrap.py	423	CODE
LOW	headroom/cli/wrap.py	2992	CODE
LOW	headroom/cli/wrap.py	3501	CODE
LOW	headroom/observability/metrics.py	324	CODE
LOW	headroom/transforms/smart_crusher.py	255	CODE
LOW	headroom/transforms/content_router.py	4278	CODE
LOW	headroom/telemetry/collector.py	101	CODE

Dead Code9 hits · 18 pts

Severity	File	Line	Context
MEDIUM	tests/test_openai_responses_context_compaction.py	357	CODE
MEDIUM	tests/test_openai_responses_context_compaction.py	358	CODE
MEDIUM	tests/test_openai_responses_context_compaction.py	369	CODE
MEDIUM	tests/test_openai_responses_context_compaction.py	373	CODE
MEDIUM	tests/test_openai_responses_context_compaction.py	378	CODE
MEDIUM	tests/test_openai_responses_context_compaction.py	381	CODE
MEDIUM	tests/test_bedrock_region.py	549	CODE
MEDIUM	tests/test_integrations/langchain/test_streaming.py	609	CODE
MEDIUM	headroom/backends/base.py	165	CODE

Overly Generic Function Names12 hits · 11 pts

Severity	File	Line	Snippet	Context
LOW⚡	tests/test_search_compressor.py	39	src/main.py:42:def process_data(items):	CODE
LOW	tests/test_search_compressor.py	24	content = """src/main.py:42:def process_data(items):	CODE
LOW	tests/test_search_compressor.py	66	C:\\Users\\dev\\src\\utils.py:20:def helper():	CODE
LOW	tests/test_search_compressor.py	176	content = """src/main.py:10:def process_data():	CODE
LOW⚡	tests/test_ccr_rust_marker_hash_bridge.py	65	original = "src/app.py:12: def handle_request(payload):\n" * 40	CODE
LOW	tests/test_compression_summary.py	101	("function handleRequest(req, res) {", " res.send('ok');", 5),	CODE
LOW⚡	tests/test_text_compressors.py	36	content = """src/main.py:42:def process_data(items):	CODE
LOW	tests/test_text_compressors.py	90	def process_data(items: list[dict]) -> dict[str, Any]:	CODE
LOW	tests/test_transforms/test_content_router.py	97	lines.append(f"src/module{i}.py:42: def process_data(input: str) -> str:")	CODE
LOW	headroom/evals/batch_compression_eval.py	716	function processData(items) {	CODE
LOW	headroom/transforms/code_compressor.py	805	... def process_data(items: List[str]) -> List[str]:	STRING
LOW	headroom/transforms/code_compressor.py	821	def process_data(items: List[str]) -> List[str]:	STRING

Cross-Language Confusion (JS/TS)1 hit · 5 pts

Severity	File	Line	Snippet	Context
HIGH	sdk/typescript/test/parity.test.ts	73	print(json.dumps({	CODE

Slop Phrases1 hit · 3 pts

Severity	File	Line	Snippet	Context
MEDIUM	.serena/project.yml	74	# Same syntax as gitignore, so you can use * and **.	COMMENT

Example Usage Blocks2 hits · 3 pts

Severity	File	Line	Snippet	Context
LOW	tests/test_proxy_count_tokens_integration.py	300	# Example usage	COMMENT
LOW	…ples/deployment/macos-launchagent/shell-integration.sh	8	# Usage:	COMMENT

TODO Padding1 hit · 1 pts

Severity	File	Line	Snippet	Context
LOW	tests/test_search_compressor.py	226	src/main.py:30:# TODO: implement later	CODE

Analysis Overview

What These Metrics Mean

Score History

Severity Breakdown

Directory Score Breakdown

Pattern Findings