kiro-validate-impl

Validates feature-level integration across completed tasks with tests, coverage, and design alignment checks.

Updated Jun 23, 2026
One-click install
npx skills add https://github.com/j5ik2o/marp-ai-base --skill kiro-validate-impl-j5ik2o
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: kiro-validate-impl
Source: https://github.com/j5ik2o/marp-ai-base/tree/main/.agents/skills/kiro-validate-impl
Command: npx skills add https://github.com/j5ik2o/marp-ai-base --skill kiro-validate-impl-j5ik2o

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? After individual tasks are implemented and reviewed, integration issues can still hide between tasks: mismatched interfaces, requirement gaps, architectural drift, and boundary violations. This Skill performs a final feature-level validation gate before declaring work complete. ## Core Features & Use Cases - Mechanical Validation: Runs the full test suite, greps for residual TODO/FIXME markers and hardcoded secrets, and performs a runtime smoke boot check. - Cross-Task Judgment Checks: Verifies interface contracts, shared state consistency, requirements coverage, design alignment, and boundary integrity across all completed tasks. - Structured GO/NO-GO Report: Produces a decision report with mechanical results, coverage mapping, ownership classification (LOCAL/UPSTREAM/UNCLEAR), and mandatory remediation steps on failure. - Use Case: After running implementation for a user-auth feature across tasks 1.1 through 1.3, invoke validation to confirm the full test suite passes, all requirements are covered, and no cross-task integration seams leak before shipping. ## Quick Start Ask the agent to validate the implemented feature, for example by saying: run kiro-validate-impl on the user-auth feature and give me the GO or NO-GO report.

Frequently Asked Questions about kiro-validate-impl

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

FAQPage Schema
How do I validate a feature after all tasks are implemented?

Invoke the validation skill with the feature name, optionally specifying task numbers. It runs the full test suite, checks requirements coverage, verifies design alignment, and returns a GO, NO-GO, or MANUAL_VERIFY_REQUIRED decision with a structured report.

What is the difference between per-task review and feature-level validation?

Per-task review checks individual task acceptance criteria during implementation. Feature-level validation catches problems only visible across tasks: integration seams, coverage gaps between tasks, architectural drift, and boundary spillover.

What happens when the test command cannot be identified?

The skill returns MANUAL_VERIFY_REQUIRED instead of GO and explains which validation command is missing. It never approves a feature when mandatory mechanical validation could not be completed.

Does the validation skill modify code or fix issues?

No, validation is read-only and idempotent, so it is safe to re-run after interruptions. On NO-GO it provides remediation steps, and you re-run implementation for targeted fixes before validating again.

When should I not rely on this validation approach?

Do not use it as a substitute for per-task review during implementation, since it deliberately skips single-task acceptance checks. It also cannot complete smoke boot verification when the required runtime environment is unavailable.