# Claude Code skills Projects at AI Tinkerers

> Canonical HTML: https://aitinkerers.org/technologies/claude-code-skills
> Markdown URL: https://aitinkerers.org/technologies/claude-code-skills.md
> Technology record last updated: 2026-02-27T15:56:06Z
> Generated: 2026-10-10T07:50:48Z

Modular, self-contained packages that transform Claude into a specialized agent: they automate complex, multi-step workflows using proven instructions and code, dynamically loading only what is needed.

Claude Code Skills are your solution for high-efficiency, specialized automation. They function as modular, self-contained packages (e.g., `SKILL.md` files) that encode domain expertise and organizational knowledge. Claude dynamically discovers and loads these resources based on context (progressive disclosure), which minimizes token usage and maximizes speed. This system allows you to execute complex, multi-step workflows—like full code review or PDF processing—with a single command, consistently applying best practices across all your projects. Skills specialize Claude: you create the expert workflow once; Claude applies it everywhere.

- Official technology site: https://docs.anthropic.com/claude/docs/skills-overview
- Public AI Tinkerers demos and talks: 3
- Result page: 1 of 1

## Recent Public Talks and Demos

### [Two agents and a lot of screenshots: rebuilding a legacy back office system](https://lausanne.aitinkerers.org/talks/rsvp_1YKm3hTMMf4)

An agent assembly line that rebuilds a live financial back-office screen by screen, after an upgrade that skipped eight years of framework versions. Around 500 screens. The planning layer is the whole project, and it is deliberately boring: no orchestration framework, no queue, no vector store, nothing outside plain Claude Code. The backlog is one JSON file behind a small CLI. Each role is a skill file in the repo. A coordinator hands five screens to a fixer subagent, then hands the same five to a validator subagent that rechecks them from scratch. Whoever made the change never approves it. State lives in the file rather than a chat context, so I kill the session mid-batch and resume a week later on the same line. Live: the progress board, the briefing an agent gets when it claims a batch, a browser pass over a screen that is broken right now, and the recorded history of one screen the validator refused.

- Event context: AI Tinkerers Lausanne September 2026 Meetup — 2026-09-29 — Lausanne
- Public talk page: https://lausanne.aitinkerers.org/talks/rsvp_1YKm3hTMMf4

### [Inside Claude Code's Brain: Debugging Claude's Behavior with Log Analysis](https://tokyo.aitinkerers.org/talks/rsvp_V2jLkUS8ltU)

Live demo of cclogviewer — an open-source MCP server and CLI tool that reads Claude Code's internal JSONL session logs and turns them into structured, queryable data: tool usage stats, session timelines, token counts, and error summaries. I'll walk through how it works, how to set it up, and how I used it to analyze 650 automated Claude Code sessions to diagnose why skills only activate ~50% of the time and how to create a logs-based improvement loop for Claude Code Subagents.

- Event context: AI Tinkerers Tokyo - Toranomon Meetup - February 19, 2026 — 2026-02-19 — Tokyo
- Public talk page: https://tokyo.aitinkerers.org/talks/rsvp_V2jLkUS8ltU

### [Claude-Reflect: Teaching Your Coding Agent Through Corrections](https://seattle.aitinkerers.org/talks/rsvp_LYdPlFYAdIc)

I will show how claude-reflect turns day-to-day corrections into persistent memory and reusable skills for Claude Code. The talk covers the two-stage pipeline (hook-based capture + semantic validation), how learnings are queued for human review, and how patterns become new /commands via /reflect-skills. I will walk through the architecture, tradeoffs in false-positive reduction, and a short live demo: correct the agent, run /reflect, and watch CLAUDE.md and skill files update.

- Event context: AI Tinkerers Seattle Meetup: Dev Tools Track — February 9th, 2026 — 2026-02-10 — Seattle
- Public talk page: https://seattle.aitinkerers.org/talks/rsvp_LYdPlFYAdIc

## Related Technologies

- [Claude Code](https://aitinkerers.org/technologies/claude-code) ([Markdown](https://aitinkerers.org/technologies/claude-code.md)) — 226 public demos
- [Python](https://aitinkerers.org/technologies/python) ([Markdown](https://aitinkerers.org/technologies/python.md)) — 673 public demos
- [cclogviewer](https://aitinkerers.org/technologies/cclogviewer) ([Markdown](https://aitinkerers.org/technologies/cclogviewer.md)) — 1 public demo
- [Claude Code CLI](https://aitinkerers.org/technologies/claude-code-cli) ([Markdown](https://aitinkerers.org/technologies/claude-code-cli.md)) — 5 public demos
- [Claude Code plugins](https://aitinkerers.org/technologies/claude-code-plugins) ([Markdown](https://aitinkerers.org/technologies/claude-code-plugins.md)) — 2 public demos
- [Claude Code skills & plugins](https://aitinkerers.org/technologies/claude-code-skills-plugins) ([Markdown](https://aitinkerers.org/technologies/claude-code-skills-plugins.md)) — 1 public demo
- [Django 6](https://aitinkerers.org/technologies/django-6) ([Markdown](https://aitinkerers.org/technologies/django-6.md)) — 1 public demo
- [MCP](https://aitinkerers.org/technologies/mcp) ([Markdown](https://aitinkerers.org/technologies/mcp.md)) — 133 public demos
- [Playwright](https://aitinkerers.org/technologies/playwright) ([Markdown](https://aitinkerers.org/technologies/playwright.md)) — 33 public demos
- [Shell](https://aitinkerers.org/technologies/shell) ([Markdown](https://aitinkerers.org/technologies/shell.md)) — 3 public demos