Posts
Memory as reliability practice
Why a good AI session should leave durable residue, and a setup skill that scaffolds a git-backed Obsidian vault to hold it. Start here; the other posts build on this vault.
memory plugin
An operating loop for AI-assisted work
Jira-backed standups, task capture, session hygiene, daily planning, and a Friday report-only AI coach: small loops that turn the vault from a filing cabinet into an operating system.
productivity plugin
Reflecting your own work back as a learning plan
A weekly engineering self-retro that reads your own sessions (or your shipped GitHub and Jira work in a cloud routine) and returns a small, evidence-cited, ranked learning plan.
learning plugin
Two Macs, one local model pool
A cautious path for sharing local inference across two Macs with NVIDIA PAIR, choosing a model that fits, and adding a loopback LiteLLM route.
local-ai plugin
Testing skills the way they actually run
Unit tests prove the checker; an eval proves the skill. A GitHub Action runs the standup skill against fixture cases and grades the transcript, kept separate from the hermetic test job.
CI workflow · skill-eval.yml
Install the plugins
Everything described in these posts is a public plugin. Add the marketplace once, then install what you want.
Claude Code
claude plugin marketplace add adzuci/claude-plugins
claude plugin install memory@adzuci-pluginsCodex
codex plugin marketplace add adzuci/claude-plugins
codex plugin add memory@adzuci-pluginsEach post ends with the install steps and prerequisites for its plugin. Tested primarily on macOS; Linux and Windows compatibility PRs are welcome.