Matyas Stoch
Community@matyasstoch · Katowice
Working with David Ondrej
Agent Skills by Matyas Stoch
Showing 43 vetted skills indexed across 2 GitHub repositories.
endless-ideas-migration
Migrates non-actionable Todoist tasks into a markdown ideas archive after priority filtering.
obsidian-vault-location
Locates the Obsidian app, vault paths, and note files on macOS.
consulting-call-questions
Manages question lists for consulting calls in Markdown files with git versioning.
vps-server-management
Manages VPS servers and on-server AI agents via SSH, deployment, and monitoring workflows.
folder-specific-claude-and-agents-md
Creates folder-scoped CLAUDE.md files with AGENTS.md symlinks for agent context.
delegating-to-agents
Orchestrates prompt delegation and polling across Pi, Codex, Claude Code, and Hermes agents.
david-sql-task-importance
Scores Todoist task importance in batches of seven and saves ratings to a SQLite database.
organize-tasks
Reviews Todoist tasks in batches of four and executes user-directed dispositions via the Todoist API.
new-society-modules
Designs New Society course module outlines through structured questioning and context file updates.
markdown-rendering
Opens markdown files in cmux right panes while avoiding the blank-render move-surface bug.
presentation-review
Builds and reviews YouTube video presentation slides using an alternating explain-showcase structure.
behavior-tracking
Logs conversation messages, decisions, and behavioral patterns to a local database.
codex-goal-loop
Writes structured goal instructions for OpenAI Codex's persistent /goal agent loop.
paid-ads-copy
Writes hooks, scripts, and copy for Meta and Google paid ad campaigns.
pre-production
Vet tools and build reproducible step-by-step guides for YouTube video recording.
improve-codebase-architecture
Identifies refactoring opportunities that turn shallow modules into deep, testable modules.
tldraw-whiteboard
Creates programmatic whiteboard graphics with the tldraw SDK and extracts content from .tldr board files.
find-skills
Searches and installs agent skills from the open skills ecosystem using the Skills CLI.
day-plan
Builds a daily priority list by asking five structured questions and organizing Todoist tasks.
cyber-audit
Audits local machine exposure to CVEs and supply-chain advisories with read-only checks.
frontend-polish
Refines existing frontend interfaces by fixing spacing, typography, color, and responsive behavior.
playwright
Automate real browsers from the terminal using playwright-cli commands.
effective-agent-skills
Guides authoring, testing, and debugging of SKILL.md agent skill files.
todoist
Manage Todoist tasks via API v1 with project filtering, pagination, and priority mapping.
Frequently Asked Questions About Matyas Stoch
FAQPage SchemaWhat tasks can I automate using Matyas Stoch's skills?▼
You can triage and migrate Todoist tasks, plan days with the Ivy Lee six-priority method, delegate prompts to coding agents in cmux panes, run DeepSWE model benchmarks, audit machines for CVE exposure, reconcile bank CSV/PDF statements, and prepare YouTube videos from idea selection through slide review.
Who are these skills designed for?▼
They are built for David Ondrej's personal setup: a solo operator running multiple coding agents (Claude Code, Codex, Pi, Hermes), producing YouTube content and New Society course modules, managing consulting calls and interviews, and maintaining local context files, Obsidian notes, and VPS servers.
How are the skills installed and triggered at runtime?▼
Skills live as SKILL.md folders distributed to four canonical locations (.claude/skills symlink, Codex, Pi at ~/.pi/agent/skills, Hermes) via the distribute-skill-to-all-agents skill, then pushed to a private GitHub repo rooted at ~/.agents. Agents load them automatically when trigger phrases match.
What are the prerequisites and dependencies?▼
Key dependencies include macOS 14+ for cmux, a running Chrome instance for CDP browser control, playwright-cli for browser scripting, the pi-web-access package for web search, an OpenRouter key for DeepSWE benchmarks, Vapi MCP access for outbound voice calls, and local SQLite databases for behavior and task scoring.
Are these skills open source and what do they cost?▼
The manifest describes a private skills GitHub repository rather than a public open-source release, so no license is specified. Runtime costs come from external services the skills call, such as OpenRouter API usage for benchmarks and Vapi for outbound phone calls.