My personal collection of Claude Code skills.
| Skill | Description |
|---|---|
| office-hours | Product Office Hours — structured product thinking partner. Two modes: Startup mode uses six forcing questions to expose demand reality, status quo, user specificity, narrowest wedge, observation gaps, and future-fit. Builder mode offers design thinking brainstorming for side projects, hackathons, learning, and open source. Produces a design doc, never code. Inspired by GStack office-hours. |
| sub-claude | Batch-process CSV/JSONL data by spawning parallel claude -p processes. Subagents are model-driven — context accumulates with each row causing cascading cost, error handling depends on the model's judgment, and you have little control over execution. For 10-1000+ rows of uniform processing (classify, translate, extract, etc.), this script-driven approach gives you isolated per-row execution (no context bleed), deterministic concurrency/retry/checkpoint behavior, and precise control over model, effort, cost budget per invocation. Includes a ready-to-use sub_claude.py. |
Symlink a skill directory into ~/.claude/skills/:
ln -s /path/to/skills/<skill-name> ~/.claude/skills/<skill-name>