Aleš Orel avatar

Aleš Orel

Community

@ulebule · Slovenija

4Followers
|
72Public Repos
|
8Published Skills

Aleš Orel publishes Claude Code skills for Azure DevOps access, SQL Server querying, code evolution, knowledge graphs, and structured decision refinement.

Skills Distribution
DomainDeveloper To...Developer Tooling .. (50%)Data Systems & Dat.. (20%)Cloud & Compute In.. (15%)AI Models & Machin.. (15%)

Agent Skills by Aleš Orel

Showing 8 vetted skills indexed across 1 GitHub repositories.

Frequently Asked Questions About Aleš Orel

FAQPage Schema
What tasks can I accomplish with ulebule's skills?

You can query Azure DevOps work items and PRs, run read-only SQL Server queries, evolve faster function implementations with OpenEvolve, build persistent knowledge graphs from codebases, and refine vague ideas or prompts through structured interviews and adversarial review.

Who are these skills designed for?

Software engineers working in .NET and Azure DevOps environments, plus developers who want rigorous decision-making: the interview-me, idea-refine, prompt-prep, and doubt-driven-development skills target anyone stress-testing assumptions before committing to plans or production code.

What are the prerequisites and dependencies?

Azure DevOps access requires the az CLI and a git remote pointing at the organization. SQL Server querying requires sqlcmd on Windows and a .NET project connection string. OpenEvolve runs through the Claude Code CLI with no API key needed.

How does the graphify skill work on a codebase?

Graphify converts code, docs, papers, images, or videos into a persistent knowledge graph with god nodes and community detection. When a graphify-out/ directory exists, codebase architecture and file-relationship questions are answered through its query, path, and explain functions first.

When should I use doubt-driven-development?

Use it when correctness matters more than speed: unfamiliar code, production changes, security-sensitive logic, or irreversible operations. It subjects every non-trivial decision to a fresh-context adversarial review before the output stands, catching errors cheaper than debugging later.