Register and run named shell commands as repeatable CI-style validation checks.
npx skills add https://github.com/SaberAloui/agri-hub --skill validation-saberaloui
Please help me install this Agent Skill. Skill: validation Source: https://github.com/SaberAloui/agri-hub/tree/main/.local/skills/validation Command: npx skills add https://github.com/SaberAloui/agri-hub --skill validation-saberaloui
npm run lint or pytest under short names (lint, test, typecheck) with upsert behavior, and list or remove them anytime. - Monitored Validation Runs: Execute one or more registered commands in a single run and receive per-command status, exit codes, durations, log file paths, and an LLM-generated failure summary with log line citations. - Run Management: Stop running validations gracefully or forcefully, and query the status of current or past runs by ID. - Use Case: After implementing a feature in a Node.js project, register lint, test, and typecheck commands, run them together, and read the generated run summary to understand exactly what failed without digging through full logs. ## Quick Start Register lint and test as validation commands for my project, then run them and tell me what failed.High-intent search queries and answers about installing and using this skill.
Register each check as a named validation command, such as lint mapped to npm run lint and test mapped to npm test. Then start a validation run with those command names to execute them together and receive structured pass/fail results.
Register both commands with setValidationCommand, then call startValidationRun with their names. The run returns per-command status, exit codes, durations, log file paths, and a generated summary explaining any failures.
Common conventions are ruff check for lint, pytest for test, and mypy for typecheck in Python; golangci-lint run, go test ./..., and go vet ./... in Go. Use short lowercase names like lint, test, and typecheck regardless of stack.
Yes, call stopValidationRun with the run ID to halt an in-progress run. It stops gracefully by default, but you can pass graceful as false to force an immediate stop.
Skip this approach for one-off test runs, where invoking the shell directly is simpler, and for deployment workflows, which belong to a deployment skill. It is designed for repeatable, named quality gates rather than ad-hoc commands.
whyte25Sets up Husky and lint-staged pre-commit hooks for JavaScript/TypeScript projects.
meta-pytorchGenerate minimal code changes to fix failing tests from /write-tests runs.
nampham2Run pytest suites, verify 100% coverage, and validate pre-commit hooks.
VektaDiagnose and resolve CI/CD pipeline failures on GitHub Actions.