improved flow and catch linting failures by buzzdan · Pull Request #8 · buzzdan/ai-coding-rules

buzzdan · 2026-01-25T13:25:00Z

This pull request introduces several improvements to the linter-driven Go development workflow, focusing on explicit skill invocation, enhanced linter-to-skill routing, stricter quality gates, and improved documentation for design and refactoring patterns. The changes ensure that skills are invoked programmatically (not just referenced), clarify how linter errors map to remediation steps, and add detailed design guidance for handling common linter failures. Additionally, the mcp__ide__getDiagnostics tool is now permitted in all relevant agents and commands, enabling better diagnostic integration.

Key changes include:

Skill Invocation and Tooling Consistency

All relevant agents, commands, and skills now explicitly allow the mcp__ide__getDiagnostics tool, ensuring consistent access to IDE diagnostics across the workflow. [1] [2] [3] [4] [5] [6] [7] [8] [9] [10]
Added <skill_invocation> sections to all major skills, requiring that skills (e.g., @testing, @refactoring) be invoked via the Skill tool, not just referenced in text. Clear tables show the mapping between notation and tool calls. [1] [2] [3]

Linter Error Routing and Design Patterns

Introduced comprehensive linter error routing tables and documentation, mapping specific linter failures to the correct skill and remediation pattern (e.g., argument-limit → options struct, function-result-limit → result type, file-length-limit → file splitting/god object decomposition). [1] [2] [3] [4]
Added detailed examples and step-by-step guidance for handling common design-related linter failures, including code snippets for before/after refactoring.

Workflow and Quality Gate Enhancements

The linter-driven development skill now enforces that all new types have corresponding test files and >0% coverage before progressing past implementation. Leaf types require 100% coverage, and warnings are issued if coverage is below 80%.
The iterative fix loop now includes orchestrator checks for "god objects" (>15 methods per type), requiring storification and service extraction before proceeding. [1] [2] [3]
The workflow is only considered complete when all extracted leaf types have tests with 100% coverage and no god objects remain.

Documentation and Usability Improvements

Updated agent and command descriptions to clarify the use of the Skill tool for invoking skills, not just referencing them. [1] [2] [3]
Added structured references and tables to help users understand linter categories, severity, and routing.

These changes collectively make the Go linter-driven workflow more robust, explicit, and maintainable by enforcing best practices and ensuring all automation is performed through well-defined skill invocations.

The rule was lost when heal-plugin removed inline patterns. Added it back to the juiciness_test section where extract-type guidance lives. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Copilot

Pull request overview

This PR tightens and documents the Go linter-driven development workflow by making skill invocation explicit, improving linter-to-skill routing guidance, and enabling IDE diagnostics across the workflow.

Changes:

Permits mcp__ide__getDiagnostics across relevant skills, agents, and commands for consistent diagnostics access.
Adds/standardizes <skill_invocation> guidance and linter routing tables to ensure skills are invoked programmatically (via the Skill tool).
Strengthens workflow quality gates and refactoring guidance (tests/coverage expectations, god-object thresholds, file-splitting patterns).

Reviewed changes

Copilot reviewed 11 out of 11 changed files in this pull request and generated 4 comments.

Show a summary per file

File	Description
go-linter-driven-development/skills/testing/SKILL.md	Updates auto-invocation phase reference for testing in the workflow.
go-linter-driven-development/skills/refactoring/SKILL.md	Adds explicit skill invocation rules, routing tables, and stronger enforcement/verification guidance.
go-linter-driven-development/skills/linter-driven-development/SKILL.md	Adds diagnostics tool permission, explicit skill invocation requirements, and expanded routing/quality-gate steps.
go-linter-driven-development/skills/code-designing/SKILL.md	Adds diagnostics tool permission and linter-triggered design patterns (options/result/file splitting).
go-linter-driven-development/commands/go-ldd-status.md	Allows diagnostics tool for status command context.
go-linter-driven-development/commands/go-ldd-review.md	Allows diagnostics tool for final review command context.
go-linter-driven-development/commands/go-ldd-quickfix.md	Allows diagnostics tool and clarifies Skill-tool-based invocation guidance.
go-linter-driven-development/commands/go-ldd-autopilot.md	Allows diagnostics tool and clarifies Skill-tool-based invocation guidance.
go-linter-driven-development/commands/go-ldd-analyze.md	Allows diagnostics tool during analysis runs.
go-linter-driven-development/agents/quality-analyzer.md	Allows diagnostics tool and documents linter categorization/routing reference.
go-linter-driven-development/agents/go-code-reviewer.md	Allows diagnostics tool and clarifies explicit Skill-tool usage for pre-commit guidance loading.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copilot · 2026-04-12T09:17:29Z

+
+<enforcement>
+Before marking refactoring complete:
+1. List all types created: `grep -r "^type.*struct" internal/`


The enforcement command grep -r "^type.*struct" internal/ assumes the project uses an internal/ directory. For Go modules that organize code under pkg/, cmd/, or the module root, this will fail and can incorrectly skip test enforcement. Consider searching from . (or using go list ./... + file scanning) instead of hard-coding internal/.

Suggested change

1. List all types created: `grep -r "^type.*struct" internal/`

1. List all types created: `grep -r "^type.*struct" .`

Copilot · 2026-04-12T09:17:29Z

+4. Coverage check: `go test -cover ./...` - leaf types must show 100%
+5. Scan for nolint in all uncommitted files (staged + unstaged):
+   ```bash
+   { git diff --name-only; git diff --cached --name-only; } | sort -u | xargs grep "//nolint" 2>/dev/null


The xargs grep "//nolint" pipeline can hang or behave unexpectedly when there are no changed files (e.g., xargs may invoke grep with no path args and wait on stdin on some platforms). Add an empty-input guard (or use an xargs form that won’t run when the list is empty) so the verification step is reliable.

Suggested change

{ git diff --name-only; git diff --cached --name-only; } | sort -u | xargs grep "//nolint" 2>/dev/null

changed_files=$({ git diff --name-only; git diff --cached --name-only; } | sort -u)

if [ -n "$changed_files" ]; then

printf '%s\n' "$changed_files" | xargs grep "//nolint" 2>/dev/null

fi

Copilot · 2026-04-12T09:17:30Z

+
+**Verification**: After refactoring, scan for nolint in all uncommitted files (staged + unstaged):
+```bash
+{ git diff --name-only; git diff --cached --name-only; } | sort -u | xargs grep "//nolint" 2>/dev/null


Same issue as above: this xargs grep "//nolint" verification can run grep with no file args when the file list is empty, which may block waiting on stdin. Add a guard / empty-input-safe approach here as well so the scan is deterministic.

Suggested change

{ git diff --name-only; git diff --cached --name-only; } | sort -u | xargs grep "//nolint" 2>/dev/null

files="$({ git diff --name-only; git diff --cached --name-only; } | sort -u)"

if [ -n "$files" ]; then

printf '%s\n' "$files" | xargs grep "//nolint" 2>/dev/null

fi

Copilot · 2026-04-12T09:17:30Z

+Verify: `grep -r "TypeName" --include="*.go" . | grep -v "type_file.go"`
+If found elsewhere → move to type's file


grep -r "TypeName" --include="*.go" . will match usages of the type (including tests, comments, and other packages), not just declarations that should be co-located. That can produce false positives and suggest moving code that doesn’t belong in the type’s file. Prefer searching specifically for declarations you want to co-locate (e.g., methods/constructors/constants tied to the type) rather than any reference to the identifier.

Suggested change

Verify: `grep -r "TypeName" --include="*.go" . | grep -v "type_file.go"`

If found elsewhere → move to type's file

Verify by searching for declarations tied to the type, not every usage of its name.

Check for receiver methods (for example, `grep -R -nE '^func \$[^)]+\\*?TypeName\$' --include="*.go" .`),

constructors like `NewTypeName`, and related `const`/`var` declarations that belong with the type.

If declaration(s) that should be co-located are found elsewhere → move those declarations to the type's file

- hooks/check-package-sizes.sh: PostToolUse hook flagging Go packages with >=13 non-test .go files (red, exit 2) or 8-12 (yellow, advisory). Portable: only find/wc/tr/sort, guards on go.mod so non-Go projects no-op. - hooks/hooks.json: registers the hook on Write|Edit|MultiEdit. - refactoring skill: new <package_decomposition> section with 3-step design review (domain naming, type scoping, split decision), Store pattern, function stutter, import direction, phased migration. Trimmed duplicate find/wc snippets now that the hook surfaces detection automatically. - pre-commit-review skill: package-size zones now detected via the hook; skill carries the HOW and a manual fallback for review-only sessions. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>

@Refactoring

- <when_to_use>: add PostToolUse hook yellow/red zone as a design-time trigger (intervene before the zone escalates). Cross-refs the full playbook in @Refactoring <package_decomposition>. - <plan_package_structure>: add the naming method (parent names the actor, sub-package names the domain object — worker/ + worker/job/), the "say pkg.Type out loud" test, anti-pattern list (handlers/, types/, common/, domain/, etc.), and the strict-downward import direction rule. - <success_criteria>: add package-name check (domain concept, not role) and import-direction check. No duplication with @Refactoring: code-designing is forward design (avoid creating red-zone packages); refactoring is reactive response (hook fires → decompose). The two link to each other. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>

buzzdan and others added 3 commits April 12, 2026 12:07

improved flow and catch linting failures

aeaf985

heal refactoring skill

2560024

restore self-validation rule in refactoring SKILL.md after heal rebase

fb03003

The rule was lost when heal-plugin removed inline patterns. Added it back to the juiciness_test section where extract-type guidance lives. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

buzzdan force-pushed the heal-plugin branch from 278ff45 to fb03003 Compare April 12, 2026 09:13

buzzdan requested a review from Copilot April 12, 2026 09:14

Copilot started reviewing on behalf of buzzdan April 12, 2026 09:15 View session

Copilot AI reviewed Apr 12, 2026

View reviewed changes

buzzdan and others added 3 commits April 21, 2026 14:31

removed lsp internal tool

c9d5486

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

improved flow and catch linting failures#8

improved flow and catch linting failures#8
buzzdan wants to merge 6 commits intomainfrom
heal-plugin

buzzdan commented Jan 25, 2026

Uh oh!

Copilot AI left a comment

Uh oh!

Copilot AI Apr 12, 2026

Uh oh!

Copilot AI Apr 12, 2026

Uh oh!

Copilot AI Apr 12, 2026

Uh oh!

Copilot AI Apr 12, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

	1. List all types created: `grep -r "^type.*struct" internal/`
	1. List all types created: `grep -r "^type.*struct" .`

-   { git diff --name-only; git diff --cached --name-only; } | sort -u | xargs grep "//nolint" 2>/dev/null
+   changed_files=$({ git diff --name-only; git diff --cached --name-only; } | sort -u)
+   if [ -n "$changed_files" ]; then
+     printf '%s\n' "$changed_files" | xargs grep "//nolint" 2>/dev/null
+   fi

-{ git diff --name-only; git diff --cached --name-only; } | sort -u | xargs grep "//nolint" 2>/dev/null
+files="$({ git diff --name-only; git diff --cached --name-only; } | sort -u)"
+if [ -n "$files" ]; then
+  printf '%s\n' "$files" | xargs grep "//nolint" 2>/dev/null
+fi

		Verify: `grep -r "TypeName" --include="*.go" . \| grep -v "type_file.go"`
		If found elsewhere → move to type's file

-Verify: `grep -r "TypeName" --include="*.go" . | grep -v "type_file.go"`
-If found elsewhere → move to type's file
+Verify by searching for declarations tied to the type, not every usage of its name.
+Check for receiver methods (for example, `grep -R -nE '^func \\([^)]+\\*?TypeName\\)' --include="*.go" .`),
+constructors like `NewTypeName`, and related `const`/`var` declarations that belong with the type.
+If declaration(s) that should be co-located are found elsewhere → move those declarations to the type's file

Conversation

buzzdan commented Jan 25, 2026

Skill Invocation and Tooling Consistency

Linter Error Routing and Design Patterns

Workflow and Quality Gate Enhancements

Documentation and Usability Improvements

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Copilot AI Apr 12, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 12, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 12, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 12, 2026

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants