Andrew Yourtchenko
Community@ayourtch
@ayourtch.bsky.social
Agent Skills by Andrew Yourtchenko
Showing 20 vetted skills indexed across 1 GitHub repositories.
using-git-worktrees
Create isolated Git worktrees with automated setup and test verification.
test-driven-development
Enforce the Red-Green-Refactor cycle by requiring a failing test before production code.
testing-anti-patterns
Prevent testing anti-patterns in unit, integration, and UI test suites.
systematic-debugging
Diagnose and eliminate root causes of bugs using a four-phase debugging framework.
sharing-skills
Contribute skills to upstream repositories via branch, commit, push, and PR workflow.
using-superpowers
Enforce disciplined workflows by invoking relevant skills before each task step.
dispatching-parallel-agents
Dispatch parallel Claude agents to investigate independent failures across test files.
executing-plans
Execute implementation plans in batches with review checkpoints.
finishing-a-development-branch
Verify tests and execute git commands to merge, push, keep, or discard a development branch.
root-cause-tracing
Trace call stacks backward to identify the original bug trigger.
condition-based-waiting
Replace arbitrary timeouts with condition polling for state changes.
brainstorming
Guide one-question-at-a-time questioning to transform ideas into validated design sections.
testing-skills-with-subagents
Apply RED-GREEN-REFACTOR cycles to diagnose and prevent skill bypass.
writing-plans
Generate zero-context implementation plans with exact file paths, code examples, and verification steps.
requesting-code-review
Dispatch a code-reviewer subagent to analyze changes against requirements.
receiving-code-review
Clarify ambiguous code-review feedback through a six-step verification workflow.
writing-skills
Automate test-driven documentation generation for SKILL.md files.
verification-before-completion
Run verification commands and confirm results before claiming completion.
subagent-driven-development
Dispatch fresh subagents per task with mandatory code reviews.
defense-in-depth
Validate data across four execution layers to prevent invalid input bugs.