# Hold My Hand

> Teach a large topic, codebase, or project as a curriculum of tiny chunks, one per turn.

- Skill: `raedmkhalaf/hold-my-hand` (Agent Skill)
- Install (CLI): `npx skillmds@latest add raedmkhalaf/hold-my-hand`
- Raw SKILL.md: https://api.skillmd.com/api/skills/raedmkhalaf/hold-my-hand/raw
- Safety review: pending (external: skill-scanner PASS, skillspector PASS)
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: RaedMKhalaf (https://skillmd.com/u/raedmkhalaf)
- Updated: 2026-09-22
- Page: https://skillmd.com/skills/raedmkhalaf/hold-my-hand

---


# hold-my-hand

Deep-research the subject silently and plan a curriculum: one sequential path of `y` chunks, each titled, ordered however teaches best. Teaching starts once `y` is fixed and every chunk has its title; the plan itself stays invisible.

Each turn shows one chunk and nothing besides:

```
📍 Chunk x of y — <title>

<2–4 tight sentences, ~45–60 words>
```

Every word teaches the subject or gets deleted. Fenced code blocks and small diagrams are free of the word budget; in late chunks prefer code over prose about code.

- Open with one calibration question ("new to it, used it, or deep in it?") unless their level is obvious. Chunk 1 is a ~50-word orientation.
- The curriculum ramps: the first chunks are prerequisite-free, internals and edge cases arrive last.
- If the subject fits in ≤3 chunks, drop the format and just explain it.
- Steering: anything or "next" → next chunk · "skip" → drop related chunks, shrink y · "deeper on X" → insert chunks, grow y · "back" → re-teach differently · a question → answer in place, x doesn't advance.