Maintainer Calibration Samples

Sample collection date: July 9, 2026.

This is a dated maintainer calibration record, not a public benchmark or normative user reference. These anonymized samples are useful for checking report volume, detector balance, and runtime on large repositories, but they should not be treated as a default-threshold mandate.

Source identities and local collection paths are intentionally omitted. Raw reports were generated outside the committed documentation set.

This pass used reproducible static settings for threshold and scoring sanity-checks: --churn off --hotspot-model static --output json --progress never.

Commands

Sample	Command	Status
large-cli-a	`reforge scan <large-cli-a> --output json --output-file target/calibration/large-cli-a.json --progress never --churn off --hotspot-model static`	Completed. Reported scan duration: 780,923 ms.
large-cli-b	`reforge scan <large-cli-b> --output json --output-file target/calibration/large-cli-b.json --progress never --churn off --hotspot-model static`	Completed. Reported scan duration: 122,485 ms.

Sample Summary

Sample	Source files	Directories	Function candidates	Raw functions	Raw types	Test files	Findings	Hotspots	Similar function groups
large-cli-a	3,205	715	7,171	38,534	7,586	737	4,844	3,126	69
large-cli-b	2,178	563	11,298	31,382	2,168	954	3,725	1,885	4

Finding severity:

Sample	Critical	Warning	Info
large-cli-a	0	4,177	667
large-cli-b	0	3,148	577

Top Finding Kinds

Kind	large-cli-a	large-cli-b
complex_function	869	793
long_function	706	611
readability_risk	555	589
many_parameters	554	86
large_file	364	157
repeated_literal	292	54
deep_nesting	291	390
debt_marker	261	94
import_heavy_file	184	40
test_duplication	122	626
large_type	65	90

Static Hotspot Summary

Raw hotspot locations are omitted from this committed note. The top static hotspots were retained only as aggregate shape data so the note documents watchlist behavior without exposing sample-specific paths.

Sample	Top hotspot priority	Top hotspot severity	File hotspots	Function hotspots	Type hotspots
large-cli-a	100	critical	2	2	1
large-cli-b	100	critical	0	4	1

Metric Percentiles

Metric	large-cli-a p50	large-cli-a p75	large-cli-a p90	large-cli-a p95	large-cli-a max	large-cli-b p50	large-cli-b p75	large-cli-b p90	large-cli-b p95	large-cli-b max
file loc	128	392	881	1,447	11,396	155	326	653	992	5,430
file imports	6	18	33	47	524	4	7	10	14	109
file public items	1	3	9	16	758	1	2	5	8	209
function loc	14	29	54	77	2,228	8	22	49	90	3,911
function complexity	1	3	7	11	444	1	1	5	10	476
function nesting depth	0	2	2	3	10	0	0	2	3	15
function parameter count	1	2	3	4	34	0	1	1	2	19
type loc	5	8	14	22	929	7	17	80	194	3,427
type member count	2	4	7	10	270	3	6	11	16	492

Observations

Both samples were scanned with churn disabled, so hotspot ranking is entirely static. This makes the sample reproducible but does not calibrate hybrid ranking behavior.
The samples are large: 1,149,253 and 652,390 total scanned file LOC respectively. Debug scans are expensive on this scale.
Neither sample produced critical findings, but both produced many critical static hotspots. This supports keeping hotspots as a watchlist rather than a hard CI gate.
large_file default 800 sits near the upper decile for both samples: large-cli-a file LOC p90 is 881 and large-cli-b p90 is 653. That threshold is broadly plausible for large mixed repositories, but stricter teams may want 600 or lower.
max_function_lines=80 is close to large-cli-a p95 at 77 and below large-cli-b p95 at 90. It is a reasonable default, but expect many warnings in CLI/TUI orchestration code.
max_function_complexity=15 is above p95 for both samples, yet complex_function is the largest finding kind in both reports. The extreme max values suggest the detector is dominated by a long tail rather than by normal functions.
max_function_parameters=5 behaves differently across samples: large-cli-a has 554 many_parameters findings while large-cli-b has 86. This should stay configurable and should not be weighted as a universal style rule.
Test maintenance signals vary by project shape. large-cli-b has 954 test files and 626 test_duplication findings, while large-cli-a has 737 test files and 122 test_duplication findings. Teams may need separate presets for production-source gates and test-maintenance audits.
Similar-function detection produced 69 groups in large-cli-a but only 4 in large-cli-b under default thresholds. The current default appears conservative for TypeScript-heavy samples but can still be costly on very large trees.

Calibration Follow-Ups

Review a stratified sample of high-priority findings with maintainers to estimate false-positive rates by detector kind.
Calibrate the documented strict, balanced, and relaxed presets against smaller libraries and service applications before changing their thresholds.
For large repositories, consider a calibration mode or docs recipe that uses --exclude-tests and higher --min-function-tokens when the goal is structural threshold calibration rather than duplication analysis.

Cross-Project Core Pass

A second pass on July 10, 2026 added smaller libraries, frameworks, a service application, and a TypeScript monorepo. Sample identities, clone locations, and raw reports remain outside the committed documentation. Each source snapshot was frozen before scanning, and all samples used the same reproducible static settings as the large CLI pass.

Sample	Shape	Source files	Function candidates	Findings	Issues	Findings merged into issues	Hotspots	Similar groups	Duration (ms)
core-rust-cli	multi-crate CLI	101	378	183	136	25.7%	99	10	6,479
core-python-framework	small framework	83	68	83	62	25.3%	40	0	486
core-go-library	small HTTP library	78	82	38	25	34.2%	41	0	381
core-js-framework	web framework	141	7	47	45	4.3%	2	0	1,076
core-java-service	service application	48	10	0	0	0.0%	0	0	96
core-ts-monorepo	plugin monorepo	1,451	556	933	556	40.4%	310	3	4,195

The pass exposed three instrumentation and model issues before the final figures above were recorded:

Python annotations and default expressions were initially traversed as if every identifier were a parameter. A representative eight-parameter function was measured as 30 parameters. Parameter extraction now counts the declared bindings only and excludes method receivers.
Java package-info.java files were initially treated as kebab-case business source, producing four naming-drift findings in the service sample. Java package and module metadata are now naming-neutral.
Dependency depth recursively enumerated simple paths in cyclic graphs. The TypeScript monorepo did not complete within 15 minutes. Depth is now computed once on the strongly connected component condensation DAG; the same complete static scan finishes in about four seconds on the calibration machine.

These corrections changed instrumentation and graph semantics, not default thresholds. The remaining large differences in report volume are therefore treated as project-shape evidence. In particular, the Java service is a useful zero-finding control for the balanced preset, while the TypeScript monorepo is a detector-balance and issue-clustering stress sample. The JavaScript framework’s happy-path test findings remain low-confidence review prompts and need maintainer labeling before any threshold or confidence change.

Keyboard shortcuts

Reforge