Verify when complete

Verify When Complete

No completion claims without fresh verification evidence.

This skill is the operational arm of ~/.claude/rules/probe-not-assume.md: claims of completion require fresh, probed evidence, not implementer self-report.

Core Principle

Evidence before claims, always. If you haven't run the verification command in this message, you cannot claim it passes.

Gate Rule

Before claiming any status or expressing satisfaction:

1. Identify: What command proves this claim?
2. Run: Execute the full command (fresh, complete)
3. Read: Full output, check exit code, count failures
4. Verify: Does output confirm the claim?
   • If NO: state actual status with evidence
   • If YES: state claim with evidence
5. Only then: make the claim

Red Flags: Stop and Verify

• Using "should", "probably", "seems to"
• Expressing satisfaction before verification ("Great!", "Done!")
• About to commit/push/PR without verification
• Relying on a previous run, not a fresh one
• Trusting subagent success reports without independent verification

Detected Toolchain

!if [ -f Taskfile.yaml ]; then echo "taskfile"; elif [ -f Makefile ]; then echo "makefile"; else echo "raw"; fi

Use the detected toolchain above to select commands from the tables below.

Verification Steps

Run each step in order. If any step fails, stop and report the error.

Format

Toolchain	Command
Taskfile	task fmt (if exists)
Makefile	make fmt (if exists)
Raw Go	go fmt ./...

Lint

Toolchain	Command
Taskfile	task lint (if exists)
Makefile	make lint (if exists)
Raw Go	go vet ./...

Build

Toolchain	Command
Taskfile	task build (if exists)
Makefile	make build (if exists)
Raw Go	go build ./...

Test

Toolchain	Command
Taskfile	task test (if exists)
Makefile	make test (if exists)
Raw Go	go test -race -count 1 -v ./...

If a specific task/target doesn't exist, fall back to the raw Go equivalent. Run tests uncached (-count 1), with race detection, no -short flag.

When to Apply

• Before committing or pushing
• Before creating or updating a pull request
• Before claiming a bug is fixed
• Before claiming a rebase or conflict resolution succeeded
• Before moving to the next task after making changes
• Before any statement asserting correctness or completion

Cross-references

• The narrowest verification command this skill produces (e.g. task test -- -run '^TestX$' -count=1) is the natural reproducer for the bisect skill. When a test that this skill once green-lit is now red, bisect with that exact command.