Skip to content
View Questi0nM4rk's full-sized avatar

Block or report Questi0nM4rk

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
Questi0nM4rk/README.md

Hey, I'm Jakub πŸ‘‹

Backend engineer (.NET) by day. Agent tooling for Claude Code by night. Measure, don't claim.

TypeScript Bun C# / .NET Shell Go to WASM Claude Code Arch Linux with Hyprland

🧭 How I build

βœ‚οΈ The best code is the code not written. Every file, flag and test has to say what is missing without it.
πŸ“ Measure, don't claim. Every number in my READMEs comes with the command that produced it.
πŸ’₯ Fail loud. One path, no silent fallback. A dead program tells no lies.
🎯 Intent first. I write down what the work is for before I write the work.
πŸ” Prove it on the real path. An end-to-end run beats a test that restates the code.
πŸ›‘οΈ Agents need rules they cannot remove. Guardrails live outside the agent: in hooks, commits and CI.

πŸ› οΈ What I build

A small stack for letting a coding agent check its own work. Each layer is its own repo and its own npm package.

flowchart LR
  A["🌳 shell-ast<br/>parse shell before it runs"] --> B["πŸͺ hook-kit<br/>agent hooks as one binary"]
  B --> C["πŸ›‘οΈ ai-guardrails<br/>rules agents can't weaken"]
  A --> C
  D["πŸ₯’ feats<br/>BDD tests for Bun"] -. tests .-> B
  D -. tests .-> C
  B --> E["🧰 qsm-marketplace<br/>Claude Code plugins"]
  F["πŸ€– cc-review<br/>Claude reviews my PRs"] -. reviews .-> C
Loading

My daily Claude Code setup as a marketplace of small plugins: skills, agents, hooks and CLI tools that help an agent see, verify and evaluate its own work.

TypeScript license

πŸ€– cc-review

A pull request reviewer that runs Claude Code in GitHub Actions and posts inline comments as its own GitHub App.

JavaScript license

πŸ›‘οΈ ai-guardrails

Hooks, a commit gate and a CI check that flag when a coding agent weakens a project's own lint and hook rules.

TypeScript license

πŸͺ hook-kit

Builds coding-agent hooks into one compiled binary that decides through stdout, stderr and the exit code.

npm license

🌳 shell-ast

A typed AST for bash, POSIX sh and mksh, so a tool can read a command before it runs.

npm license

πŸ₯’ feats

A BDD runner that turns Gherkin feature files into typed bun:test suites.

npm license

🧠 tagen

Resolves a query against a catalog of skill cards and returns the matching set as a manifest.

license

πŸ—‚οΈ vori

Queries Obsidian-style markdown vaults from the terminal: tags, wikilinks, full text.

license

πŸ§ͺ Experiment: codebase-mcp was my first try at code retrieval for agents. It taught me where over-engineering starts, and its README says so.

🎴 Away from the terminal

My split Dactyl keyboard, left half

  • πŸƒ I do cardistry and card tricks.
  • πŸŒ€ I obsess over a random new thing for a month, then forget what I did a week later.
  • πŸ”§ I like owning my tools, so I build them. Even my desktop is my own: qsm-desktop is a Quickshell shell for Hyprland, built on the qsm-shell component library and previewed in qsm-workbench.
  • ⌨️ I type on a split Dactyl I built myself, running ZMK on nice!nano v2 boards: zmk-config-lozor.

Popular repositories Loading

  1. GameHacking GameHacking Public archive

    C++ 1

  2. claude-implementation-partner claude-implementation-partner Public archive

    🧠 SuperClaude-style implementation advisor for .NET developers - Smart brainstorming partner with memory, auto-activating agents, and essential integrations

    Shell 1 1

  3. codeagent codeagent Public archive

    Research-backed autonomous coding framework for Claude Code. Skills, commands, and memory infrastructure for accuracy-optimized development.

    Shell 1

  4. amem-mcp amem-mcp Public archive

    A-MEM brain-like memory MCP for Claude Code - semantic memory with auto-linking

    Python 1

  5. shell-ast shell-ast Public

    Typed AST for bash, POSIX sh and mksh in TypeScript. mvdan/sh compiled to WASM. For tools that read commands before they run.

    TypeScript 1 1

  6. hook-kit hook-kit Public

    TypeScript library that compiles command-gating rules into one hk binary used in place of bash -c. Parses with a real shell parser. No model calls.

    TypeScript 1