Skill Locker
All skills
v1.3.03 loop iterations

Concept Mapper

Linear text hides structure. A textbook chapter presents concepts in sequence and that implied order is usually a lie — most domains are networks, hierarchies, or webs of tension, not lists. The Concept Mapper makes the actual shape of a topic visible.

What this skill does

A summary of a topic is a list. A concept map is a structure. The difference matters because understanding a field deeply isn't about remembering what's in it — it's about knowing how the parts connect, which ones are foundational and which are derived, where the tensions live, and which concepts are quietly the same idea under different names. That structural knowledge is what separates someone who has memorised a field from someone who can reason inside it.

The skill picks the right map type for the topic. Hierarchical trees for taxonomies and classification systems. Network maps with labelled edges for many-to-many domains like philosophy or systems thinking. Process maps for sequences and decision trees. Layered stacks for topics with clear foundational-to-advanced progression. Tension maps for trade-off-driven domains — speed versus quality, depth versus breadth, the polarities where real strategic thinking happens. Forcing a network into a tree loses information. Forcing a hierarchy into a web hides it. The map type is part of the deliverable.

Three rules keep the output usable. 8 to 20 concepts — more than that and the map becomes a ball of yarn that signals everything connects to everything, which is never true. Every edge labelled — "A relates to B" is barely better than nothing; "A enables B", "A contradicts B", "A is a special case of B" is where the understanding lives. Indicate what's foundational — every map names the two or three concepts that, if you understand nothing else, you should understand these.

Output comes in the format that fits the user. Mermaid diagrams for Obsidian and GitHub. Markdown outlines for portability across any tool. Structured lists with indentation and arrows when the relationships are too dense for a clean diagram. Each map ships with a reading guide: where to start, the critical learning path through the concepts, one hidden connection that isn't obvious, and an honest note on what the map intentionally leaves out and where to go deeper. For Obsidian users, every concept comes pre-formatted as an atomic-note title with suggested [[wikilinks]] — ready to seed a new cluster in the vault.

When this triggers

  • ·You're starting a new field and need the lay of the land before you read another article
  • ·You've read four books on a topic and the ideas still feel disconnected
  • ·You're teaching a topic and need to know the order concepts actually depend on each other
  • ·You want a Mermaid diagram you can paste into Obsidian to seed a new note cluster
  • ·You're making a decision and need to see how the relevant concepts trade off against each other

Example

Trigger

User: 'Map machine learning for me — I'm a backend engineer learning it for the first time.'

Output

Scope: ML fundamentals for someone who already codes. 14 concepts. Map type: layered + network hybrid. Layer 1 (Foundation): Probability, Linear algebra, Optimisation Layer 2 (Core methods): Supervised learning, Unsupervised, RL Layer 3 (Applied): Loss functions, Gradient descent, Regularisation Layer 4 (Modern): Neural networks, Transformers, Embeddings [Mermaid diagram with labelled edges — "enables", "is a type of", "in tension with"] Reading the map: · Start here: Probability + Linear algebra (everything depends on these) · Critical path: 1 → Loss functions → Gradient descent → 2 → 3 · Hidden connection: Regularisation and Bayesian priors are the same idea wearing different clothes · What's missing: deployment, MLOps, evaluation in production — worth a separate map Obsidian seeds: 14 atomic note titles with suggested [[wikilinks]].

Get this skill + 7 more

Get the full Learning & Growth pillar (8 skills) or the complete library.

Get the full stack — $299

What you get

  • 198-line SKILL.md, ready to drop into ~/.claude/skills/
  • Tested through 3 Karpathy-loop iterations (versions v1.0.0 → v1.3.0)
  • Triggers automatically when relevant — no command to remember
  • Lifetime updates as the skill is refined further

More from Learning & Growth

Browse the full library

297 skills across 31 categories. One purchase, lifetime updates.

See all bundles