iVintik

ivintik

AI-driven game testing with TITAN-inspired agent architecture. Automates Unity game testing via HTTP Bridge SDK — state inspection, action scripting, coverage tracking, visual regression, and autonomous exploration.

Plugin lifecycle management — issue tracking, health audits, optimization, fixing, releasing, and marketplace operations for Claude Code plugins

Setup and project init for Claude Code tools — Context7, Serena, Beads, Agent Mail, Atlas, Relay, Codeman

Project registry & awareness across all Claude sessions — discover, cache, and serve project metadata via MCP tools

Autonomous development toolkit — autopilot, quality gates, story validation, autonomy assessment

Generic gamedev side-tooling — sound (freesound fetch, manifest-driven slicing, normalize) and image (info, chroma key, trim, sprite-sheet slice/pack, AI sprite-gen via banana). Category-first layout, CLI-driven, no MCP.

Generic personal toolkit. v0.1.0 ships the secrets category — dotenvx-based secret management with hierarchical .env resolution (home → project), single-master-key model, JSON export/import for vault round-tripping, and a one-shot keychain → dotenvx migrator.

Cross-project issue routing & work handoff — route issues to GitHub/GitLab/Jira, transfer work between sessions and people

Makes autonomous coding agents produce software that actually works — real-world verification, observability, and mechanical enforcement via Claude Code hooks.

Wraps the ai-pkm CLI as a Claude Code plugin — query, ingest, lint, and inspect a personal knowledge wiki via slash commands and an auto-triggering skill.

Attach Claude Code to a live audio meeting transcript — tools, curated /mcp__steno__* prompts, subscribable transcript resource, and the experimental claude/channel capability for ambient push reactions to committed utterances. Pairs with the steno macOS daemon.