verification-loop

Validates completed work through targeted tests, lint, type checks, and manual review.

Updated Dec 1, 2025
One-click install
npx skills add https://github.com/Boulea7/ohmyclaude --skill verification-loop-boulea7
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: verification-loop
Source: https://github.com/Boulea7/ohmyclaude/tree/main/src/ohmyclaude/templates/skills/verification-loop
Command: npx skills add https://github.com/Boulea7/ohmyclaude --skill verification-loop-boulea7

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Work is often marked complete without adequate validation, leading to undetected regressions and broken builds. This Skill enforces a disciplined verification loop so every change is checked against the smallest meaningful set of tests and checks before being declared done. ## Core Features & Use Cases - Risk-Based Validation Ordering: Runs fast targeted checks first, then escalates to broader lint, type, build, and packaging checks as risk grows. - Explicit Verification Reporting: Compares expected versus actual results and reports exactly what was verified and what was not. - Honesty Guardrails: Prevents claiming verification that was never run and forces explicit statements of blockers and skipped checks. - Use Case: After refactoring a payment module, run the verification loop to execute the module's unit tests, then lint and type checks, and finally a manual review of the high-risk transaction paths before merging. ## Quick Start Before marking this task complete, run the verification loop to validate my changes with the appropriate tests and checks.

Frequently Asked Questions about verification-loop

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I verify code changes before marking a task complete?▼

Run the smallest meaningful validation set first, such as targeted unit or integration tests for the changed behavior. Then escalate to lint, type checks, and build checks as the risk or blast radius of the change increases.

What order should tests and checks run in a verification workflow?▼

Start with targeted unit or integration tests, then run lint and type checks, followed by build or packaging checks. Finish with a manual review of high-risk paths to catch issues automated checks miss.

When should I escalate from unit tests to broader checks?▼

Escalate when the risk or blast radius of the change increases, such as modifications touching shared modules or critical paths. Broader checks like builds and packaging validation catch integration-level failures.

What should I do if some verification checks cannot be run?▼

State the blockers and skipped checks explicitly in your report rather than claiming full verification. The guardrails require never claiming verification you did not actually run.

What are the limitations of a manual verification loop?▼

The loop depends on the quality of the chosen validation set, so a poorly scoped test selection can miss defects. It also relies on manual review for high-risk paths, which is slower than fully automated pipelines.