Verify

Overview

verify turns vague "it should work" claims into concrete evidence. Use it when you want confidence that a feature, fix, or refactor actually works before declaring completion.

This skill is the lightweight counterpart to the verifier agent — it's a focused workflow you can invoke directly without spawning a subagent.

Usage

/oh-my-claudecode:verify

State what behavior must be proven. The skill walks through verification options in priority order.

Workflow

1. Identify the exact behavior that must be proven
2. Prefer existing tests first
3. If coverage is missing, run the narrowest direct verification commands available
4. If direct automation is not enough, describe manual validation steps and gather concrete observable evidence
5. Report only what was actually verified

Verification Order

1. Existing tests (npm test, pytest, etc.)
2. Typecheck / build (tsc, npm run build)
3. Narrow direct command checks (curl, node -e ..., etc.)
4. Manual or interactive validation

Rules

• Do not say a change is complete without evidence
• If a check fails, include the failure clearly
• If no realistic verification path exists, say that explicitly instead of bluffing
• Prefer concise evidence summaries over noisy logs

Output Format

Verified:           <what was proven>
Commands run:       <what was executed>
Passed:             <what passed>
Failed/unverified:  <what didn't pass or wasn't checked>

Related Skills

• debug — diagnose runtime issues
• trace — evidence-driven causal tracing
• ultraqa — full QA cycling loop

Reference