multi-agent-e2e-validation_skill

This skill enables parallel end-to-end validation of database refactors by coordinating multi-agent tests and producing structured findings to ensure release

TypeScript

14

GitHub Stars

1

Bundled Files

2 months ago

Catalog Refreshed

4 months ago

First Indexed

Readme & install

Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.

Installation

Preview and clipboard use veilstrat where the catalogue uses aiagentskills.

npx veilstrat add skill terrylica/cc-skills --skill multi-agent-e2e-validation

SKILL.md11.2 KB

Overview

This skill runs a prescriptive multi-agent workflow to perform parallel end-to-end validation for database refactors and schema migrations. It spawns environment, data-flow, and query-interface agents to find functional, performance, and compatibility issues before release. The output includes structured findings, artifacts, and a go/no-go decision based on classified bug severity.

How this skill works

The workflow creates three layered agents: Environment Setup (prerequisite), Bulk Loader (data ingest and performance), and Query Interface (query correctness and edge cases). Environment validation runs first; data-flow and query agents run in parallel after the environment passes. Each agent emits test scripts, logs, artifacts, and a findings report that feed into a centralized VALIDATION_FINDINGS.md and the release decision framework.

When to use it

Validating major database refactors (e.g., migrating storage engines or DB vendors)
Testing schema migrations and type conversions before production rollout
Benchmarking bulk ingestion pipelines and deduplication logic
Pre-release verification for database-centric systems with strict SLOs
Detecting query compatibility issues across DB engines

Best practices

Always run the Environment agent first to catch infra or schema issues early
Execute Bulk Loader and Query Interface agents in parallel when safe to speed coverage
Produce structured test outputs (name, pass/fail, details) and archive artifacts
Classify findings by severity and update VALIDATION_FINDINGS.md incrementally
Re-run failing tests after fixes and record verification evidence

Example use cases

Refactor from file-based storage to QuestDB and validate ingestion, types, and queries
Validate a multi-month bulk ingestion pipeline with deduplication and re-ingestion scenarios
Test query APIs (get_latest, get_range, execute_sql) for boundary and error cases
Benchmark ingestion throughput against SLOs and flag performance regressions
Run a pre-release decision using Go/No-Go criteria driven by classified bugs

FAQ

Stop further validation. Fix environment or schema issues, then re-run Environment before starting data-flow or query agents.

How are bugs prioritized for release decisions?

Bugs are classified as Critical/Medium/Low. Any unfixed Critical blocks release. Ship requires all Critical fixed; Mediums may be accepted per policy.