[refactor] pkg/cli helper outliers, split host classification, and codemod frontmatter boilerplate

[github-actions]

🔧 Semantic Function Clustering Analysis

Analysis of github/gh-aw — pkg/ Go sources (non-test). Detection: Serena/gopls-assisted symbol analysis + naming-pattern clustering + ripgrep cross-file verification. Every finding below was confirmed by reading the actual code.

Overview

Metric	Value
Non-test .go files in pkg/	981
Primary packages analyzed	pkg/workflow (420 files), pkg/cli (338 files)
Confirmed refactoring findings	6
Verified non-issues (already well-organized)	3

Headline: pkg/workflow is already strongly consolidated (dedicated validation_helpers.go, map_helpers.go, allocation_helpers.go, parse_helpers.go; _wasm.go "duplicates" are intentional build-tag stubs). The actionable opportunities concentrate in pkg/cli: a few generic helpers living in feature-specific files, two split host-classification functions, and a wide spread of frontmatter type-assertion boilerplate across codemods.

Critical / high-confidence findings

1. Outlier helper — isGHCLIAvailable() in a download-specific file

• Defined: pkg/cli/agent_download.go:11 — func isGHCLIAvailable() bool
• Used by 4 unrelated files: pr_helpers.go:18, add_command.go:234, enable.go:72, run_workflow_execution.go:91
• A generic "is the gh CLI on PATH" check trapped in a file about agent downloads.
• Recommendation: move to pkg/cli/git.go (already the home of CLI-availability checks like isGitRepo()), or a small cli_env.go.

2. Outlier helper — getParentDir() reimplements path logic

• Defined: pkg/cli/helpers.go:16 — func getParentDir(path string) string (hand-rolled strings.LastIndex(path, "/"))
• Used in resources.go:101, includes.go:99,154, dispatch.go:116,316
• Pure string/path manipulation sitting among workflow-domain helpers; it's a weaker, slash-only variant of path.Dir/filepath.ToSlash+Dir.
• Recommendation: replace call sites with path.Dir (paths here are already forward-slash workflow paths), or relocate to pkg/fileutil/pkg/stringutil.

Medium findings

3. Split host-classification logic — isGitHubHost() vs isGHESHost()

• pkg/cli/spec.go:238 — func isGitHubHost(host string) bool (matches github.com, raw.githubusercontent.com, *.ghe.com, *.github.com)
• pkg/cli/init.go:315 — func isGHESHost(host string) bool (strips port, excludes github.com and *.ghe.com)

Both encode the same domain knowledge (github.com, the *.ghe.com cloud-tenant suffix) but as complementary predicates in two different files, each with a single caller. A future change to GitHub host rules must touch both, independently.

Recommendation: co-locate as a small host-classification helper (e.g. cli/github_host.go) with clearly-named isGitHubCloudHost / isGitHubEnterpriseHost, so the domain knowledge lives in one place.

4. Frontmatter map type-assertion boilerplate across ~40 codemods (scattered pattern)

Across pkg/cli/codemod_*.go there are 96 raw type-assertion sites (.(map[string]any), .(string), .(bool)) spread over ~40 files, repeatedly destructuring frontmatter map[string]any by hand, e.g. in codemod_serena_import.go (7), codemod_difc_proxy.go (5), codemod_github_app.go (5), codemod_copilot_requests_feature.go (5), codemod_workflow_dispatch_required.go (5).

pkg/workflow already solved the equivalent problem with parse_helpers.go/map_helpers.go. pkg/cli has no equivalent.

Recommendation: add pkg/cli/frontmatter_helpers.go with generic typed getters:

• getStringField(fm map[string]any, path ...string) (string, bool)
• getBoolField(fm map[string]any, path ...string) (bool, bool)
• getMapField(fm map[string]any, key string) (map[string]any, bool)

This removes dozens of fragile, repeated , ok := assertion blocks. (Mechanical, incremental — migrate one codemod at a time.)

Low / borderline findings

5. Permission-error formatters share a shape (pkg/workflow)

• pkg/workflow/dangerous_permissions_validation.go:79 — formatDangerousPermissionsError(writePermissions []PermissionScope) error
• pkg/workflow/github_app_permissions_validation.go:95 — formatWriteOnAppScopesError(scopes []PermissionScope) error
• pkg/workflow/github_app_permissions_validation.go:229 — formatGitHubAppRequiredError(appOnlyScopes []PermissionScope) error

All three follow build []string lines → strings.Join("\n") → errors.New(...) over a []PermissionScope. The per-scope bullet rendering (fmt.Sprintf(" - %s: ...", scope)) is duplicated.

Recommendation (optional): extract one formatScopeList(scopes []PermissionScope) []string (or formatPermissionError(title, body string, scopes []PermissionScope) error). Low priority — the wrapping messages are genuinely context-specific.

6. sortedRemainingPermissionKeys() is a thin redundant wrapper

pkg/cli/codemod_dependabot_permissions.go:242 is just return sliceutil.SortedKeys(remaining). Inline it at the (single) call site or drop the wrapper. Trivial.

Verified non-issues (no action)

• pkg/workflow _wasm.go "duplicates" (docker_validation_wasm.go, git_helpers_wasm.go, github_cli_wasm.go) are intentional //go:build js || wasm stubs — correct Go idiom, not duplication.
• Codemod sorted-key helpers already delegate to pkg/sliceutil.SortedKeys[K cmp.Ordered, V any] (sliceutil.go:63) — the central util is already in use (only the wrapper in finding 6 is redundant).
• pkg/workflow helper centralization is already mature (validation_helpers.go documents 70+ extracted patterns) — no scatter found there.

Suggested order of work

1. (P1, ~30 min) Move isGHCLIAvailable() → git.go; replace/relocate getParentDir(). Pure relocation, low risk.
2. (P2, ~1 h) Consolidate isGitHubHost/isGHESHost into one host-classification file.
3. (P2, incremental) Add frontmatter_helpers.go typed getters; migrate codemods opportunistically.
4. (P3, optional) Extract permission-error scope-list formatter; inline the redundant sorted-keys wrapper.

Analysis metadata

• Files scanned: 981 non-test .go under pkg/; deep analysis on pkg/workflow + pkg/cli (77% of files).
• Method: Serena/gopls symbol overview + naming-pattern clustering (parse*/validate*/format*/sorted*/is*Host) + ripgrep cross-file confirmation; each finding read-verified.
• Analysis date: 2026-06-30.

Generated by 🔧 Semantic Function Refactoring · 240.5 AIC · ⌖ 15 AIC · ⊞ 9.2K · ◷

• expires on Jul 1, 2026, 4:15 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 · 3.06 AIC · ⌖ 15.1 AIC · ⊞ 6K · ◷