[refactor] Semantic function clustering: decompose large multi-cluster files (audit.go, import_field_extractor.go)

[github-actions]

🔧 Semantic Function Clustering Analysis

Analysis of repository: github/gh-aw — non-test Go sources under pkg/

Overview

Analyzed 1,011 non-test .go files. Function names were extracted across the tree, clustered by naming/purpose, and cross-file name collisions were verified against their implementations to distinguish real duplication from legitimate Go patterns.

Headline: the codebase is already well-factored. Shared utilities live in dedicated packages (stringutil, sliceutil, setutil, fileutil, jsonutil, timeutil, semverutil, ...), and every naming-based duplicate candidate turned out to be a legitimate pattern (see Verified non-findings). The only actionable refactoring signal is a small number of very large, multi-cluster files that mix several internal function groups and could be split by cluster.

Package	Non-test files	Note
pkg/workflow	431	Compiler / YAML generation
pkg/cli	339	Cobra commands
pkg/parser	43	Frontmatter / imports
pkg/console	29	Rendering
others	~169	util + linters

Key finding: large multi-cluster files (one-file-per-feature candidates)

These files are internally on-topic but large enough to contain distinct function clusters that would read better as separate files. Function names below are verbatim from source.

pkg/cli/audit.go — 1,257 lines, 63 functions

Four separable clusters live in one file:

• Command setup / orchestration: NewAuditCommand, registerAuditCommandFlags, runAuditCommand, runAuditSingle, runAuditMulti, resolveAuditCommandArgs
• Analysis launching: launchCoreAuditAnalyses, launchMetricsAnalysis, launchJobDetailsAnalysis, launchFirewallAuditAnalyses, launchSupplementalAuditAnalyses, collectAuditAnalysisResults
• Rendering: renderAuditReport, renderAuditOutput, renderAuditGatewayMetrics, renderAuditUnifiedTimeline, renderAuditCompletion, renderAuditJobSummary
• Job / step extraction: fetchAuditJobLog, extractAuditJobDetails, extractRequestedStepOutput, extractFirstFailingStepOutput, findFirstFailingStep

Likely outliers (workflow-metadata helpers, not audit-specific): fetchWorkflowRunMetadata, buildWorkflowRunMetadataArgs, classifyWorkflowRunMetadataError, resolveWorkflowRunDisplayName, resolveWorkflowDisplayName, extractWorkflowNameFromYAML.

Suggestion: split rendering into audit_render.go (peers audit_report_render.go, audit_diff_render.go already exist), job extraction into audit_job.go, and move the workflow-metadata helpers next to existing workflow-metadata code.

pkg/parser/import_field_extractor.go — 1,369 lines, 57 functions

Three clusters:

• Field extraction (core): extractAllImportFields, extractEngineConfig, extractConfigFields, extractActivationFields, extractStepAndJobFields, appendJSONBuilderField, appendYAMLBuilderField, ...
• Schema validation: validateGitHubAppJSON, validateWithImportSchema, validateObjectInput, validateImportInputType, applyImportSchemaDefaultsFromFrontmatter
• Input substitution: substituteImportInputsInContent, buildImportInputReplaceFunc, resolveImportInputPath, resolveImportInputValue, computeImportRelPath

Suggestion: extract into import_validation.go and import_input_substitution.go, leaving the extractor focused on field extraction.

Other size outliers (verify before splitting)

File	Lines	Funcs	Character
pkg/workflow/safe_outputs_config.go	1,254	8	Data-heavy (config structs) — not a clustering target
pkg/workflow/compiler_yaml_main_job.go	1,163	20	Large template emitters
pkg/workflow/frontmatter_extraction_yaml.go	1,179	16	YAML extraction
pkg/workflow/compiler_yaml.go	1,129	18	YAML generation
pkg/cli/logs_orchestrator.go	1,278	13	Log orchestration
pkg/workflow/compiler_activation_job_builder.go	1,110	43	Cohesive — all *Activation* step-builders; borderline (permissions sub-cluster buildActivation*Permissions + env sub-cluster buildActivation*Env could be extracted)
pkg/workflow/awf_helpers.go	1,054	24	Helper grab-bag — worth a cluster review
pkg/workflow/domains.go	1,061	—	Mostly domain data

pkg/parser/remote_fetch.go (1,553 lines) is already tracked for splitting — see #43792 and #43834 — and is excluded here to avoid duplication.

Verified non-findings (transparency)

Naming-based duplicate detection produced candidates that were checked against source and are legitimate — no action needed:

• SanitizeName in both pkg/workflow/strings.go and pkg/stringutil/sanitize.go — the workflow copy is a thin re-export (return stringutil.SanitizeName(...)) with a SanitizeOptions type alias. Intentional, already consolidated.
• IsStdoutTerminal / IsStderrTerminal in tty.go + tty_wasm.go — build-tag variants (wasm), required.
• RenderMCPConfig, GetSupportedEnvVarKeys, GetInstallationSteps, GetExecutionSteps, ... (×9 each) — engine interface implementations (polymorphism), not duplicates.
• Error / String / UnmarshalJSON / Unwrap — standard-library interface implementations.
• contains, findRepoRoot collisions — all in _test.go files (excluded from scope).

Next actions

• Priority 1 (low risk, high clarity): split pkg/cli/audit.go and pkg/parser/import_field_extractor.go by the clusters above; relocate the workflow-metadata outliers out of audit.go. Pure function moves within a package — no signature or behavior change; existing tests should pass unchanged.
• Priority 2: review pkg/workflow/awf_helpers.go (24-function grab-bag) for cluster extraction; consider extracting the permissions/env sub-clusters from compiler_activation_job_builder.go.
• Skip: safe_outputs_config.go / domains.go (data-heavy, not function clusters); remote_fetch.go (already tracked in [file-diet] Refactor pkg/parser/remote_fetch.go (1553 lines) into focused modules #43792 / refactor(parser): split remote_fetch.go (1553 lines) into focused modules #43834).

Analysis metadata

• Non-test Go files analyzed: 1,011
• Function clusters reviewed: naming-pattern + cross-file collision analysis
• Real duplicates found: 0 (all candidates verified legitimate)
• Outlier / god-file split candidates: 2 high-confidence (audit.go, import_field_extractor.go) + follow-ups
• Method: function-signature extraction + Serena-configured Go workspace + source verification of every collision
• Analysis date: 2026-07-07

Generated by 🔧 Semantic Function Refactoring · 256.6 AIC · ⌖ 15.4 AIC · ⊞ 9.2K · ◷

• expires on Jul 8, 2026, 4:14 PM UTC-08:00

[github-actions]

🍪 Issue Monster selected this for Copilot

I've identified this issue as a good candidate for automated resolution and requested assignment to the Copilot coding agent.

If assignment succeeds, the Copilot coding agent will analyze the issue and create a pull request with the fix.

Om nom nom! 🍪

🍪 Om nom nom by Issue Monster · 1.75 AIC · ⌖ 3.98 AIC · ⊞ 6.5K · ◷