code-quality-gate

name	code-quality-gate
description	Build, test, format, and lint requirements that must pass before code review. Load when checking implementation completeness or running the quality gate.
compatibility	["claude-code","opencode","github-copilot"]
metadata	{"version":"1.0","author":"team"}

Quality Gate

Before invoking reviewers, all checks must pass. Run make ci to execute the full pipeline.

Required Checks

Check	Command	What It Verifies
Tidy	go mod tidy	Dependencies are clean
Format	go fmt ./...	Code is formatted
Vet	go vet ./...	Common mistakes caught
Lint	make lint	golangci-lint rules pass
Deps	make deps-check	No prohibited dependencies
Test	go test ./...	All tests pass
Build	go build -o bin/reference	Binary compiles
Autofix audit	— (procedure below)	Every design-doc-autofix record stays within bounds; every uncommitted change to a design-doc path is covered by a design-doc-autofix or design-block record since last commit.

Autofix Audit Procedure

Run this before declaring the gate passed. The audit enforces the protocol in review-checklist § Root-Applied Autofix on Design Docs. No script — the coordinator runs the checks the same way it follows the validation gates in pipeline-handoff.

Step 1 — Static re-validation of every autofix record. Read .scratch/handoff.jsonl. For each record where type == "design-doc-autofix", verify:

Check	Rule
File scope	file matches one of the design-doc paths declared in doc-review § Autofix on Design-Doc Paths (docs/system-design.md or any docs/adr/*.md).
Category	category is writing-standards or structural. Any other value fails.
Size bounds	lines_changed ≤ 5 AND chars_changed ≤ 200.
No heading touch	Neither old_content nor new_content contains a ## line.
No anchor change	<a id="..."> values in old_content are identical to those in new_content.
No REQ-ID change	REQ-ID tokens (regex REQ-[A-Z]+-\d{3}) in old_content are identical to those in new_content.
No code-fence touch	Neither old_content nor new_content contains a ``` line.
No link-target change	Markdown link targets (the URL inside ](...)) in old_content are identical to those in new_content.
Verbatim fix	new_content equals source_finding.fix byte-for-byte.

Any failure: do NOT declare gate-pass. Append a review-feedback finding under your own coordinator role naming the failing record by handoff.jsonl line number, then route to system-design-expert with the failing record and instructions to revert the autofix and either redo it correctly or convert it to a substantive design-block.

Step 2 — Direct-edit detection. Run git diff --name-only HEAD -- docs/system-design.md docs/adr/. For each path returned:

• Read .scratch/handoff.jsonl. Confirm at least one of the following exists with ts later than the last commit's timestamp (git log -1 --format=%cI):
  • A design-doc-autofix record whose file equals the path, or
  • A design-block record listing the path in primary_paths or supporting_paths.
• If a path is dirty but no covering record exists, the gate fails. The change was made outside the protocol — revert it or write the missing record before re-running the gate.

Step 3 — Result. Only declare the autofix-audit check green when Steps 1 and 2 both pass. Record the outcome alongside the other quality-gate results.

Optional Checks

Check	Command	When Required
Race detector	go test -race ./...	When concurrency is involved (requires gcc)
Container build	make podman-build	When project uses containers

Completion Criteria

A feature is complete when:

• All TDD cycles finished
• Self-review pass complete (see tdd-workflow § Self-Review Pass — a clause walk, not a record)
• All tests pass (go test ./...)
• Code formatted (go fmt ./...)
• Lint passes (make lint)
• Dependency policy passes (make deps-check)
• Autofix audit passes (see "Autofix Audit Procedure" above)
• Config example reflects any new/changed config fields (if applicable)
• All four reviewers approve
• No pending escalations (or human approved)

Stop at done

Once every box above is checked, stop. Polish past the bar — extra refactors, additional tests for the same behavior, prose tightening on a passing PR — spends tokens without raising quality and is explicitly out of scope. The eight-clause bar is defined across docs/tdd-principles.md, docs/testing-principles.md, and docs/ddd-principles.md, with the canonical slug list in review-checklist § Quality-Bar Clause Mapping; if the diff meets the eight clauses, the work is done.