Skip to content

krzysztofdudek

6
plugins
6
skills
3
servers

Mine an architecture graph out of a repository that has none: grain propose reads the code and the full git history and writes node types, nodes, relations, cycles and the rules the repository's own code can prove, each one carrying the evidence that produced it and the count of places that already break it, as a tree Yggdrasil loads and accepts with one command. The agent-facing questions it has always answered still work from the same model: where <intent> (where such things live, what is expected there, exemplars to copy), what <term>, how <query>, check <file> (how a change sits against the local norm), review, completeness, status/report/rules, export, decide steer|boundary|waive — plus hooks that speak unbidden and an MCP server exposing every command as a tool to any MCP client. Statistical answers mined from syntax trees and git history; no model calls, no network, never blocks.

1 skill · 1 servercoding-agentsclaude-code-plugincontext-engineering

Run a mission too big for one agent: turns your coding agent into a director that raises a horde — workers in worktrees, one ticket each, a one-shot review of every change that can send work back but never approve it, an architect subagent with a veto over the graph, legislate that writes each territory's law down, and a retrospective that closes the mission out.

1 skill · 1 serverclaude-codemulti-agentsubagents

Skill that turns your coding agent into the director of a feature branch: an issue loop (kept out of git by default, committed when you ask) where everything seen becomes an issue, a worker per issue works in its own worktree, evidence decides what merges, and the loop leaves no trace on main.

1 skill · 1 serverclaude-codemulti-agentcoding-agents

A communication layer for your AI coding agent: plain language, no jargon, no invented timelines, and one clear yes before anything irreversible — shaping how the agent talks to you, not how it works.

1 skillclaude-codeai-agentsclaude-code-plugin

Autonomous experimentation skill — your AI coding agent designs experiments, tests hypotheses, discards failures, keeps wins. Runs overnight while you sleep.

1 skillclaude-coderesearchcoding-agents

Skill that makes your coding agent stop and ask instead of silently resolving spec ambiguity, picking up workarounds, or shipping 'close enough'. Accuracy over autonomy.

1 skillclaude-codecoding-agentsclaude-code-plugin