Skills: Bundled Procedural Knowledge

Skills are reusable bundles of instructions plus optional scripts and assets. They're how Claude Code learns a procedure once and reapplies it everywhere.

9 min · Reviewed 2026

What a skill is

A skill is a folder with a SKILL.md (the instructions), optional scripts (helpers the model can call), and optional assets (templates, examples). When the model recognizes that a user request matches the skill's description, it loads the skill and follows its instructions instead of improvising.

Where skills live

User-scoped: in `~/.claude/skills/<skill-name>/` — available across all projects
Project-scoped: in `.claude/skills/<skill-name>/` inside a repo — versioned and shared with the team
Plugin-scoped: shipped by official or third-party plugins, prefixed with the plugin name
Discoverable: the SKILL.md frontmatter description is what triggers the skill

Anatomy of a skill

--- name: release-notes description: Generate user-facing release notes from recent commits. Triggers when the user asks for release notes, changelog, or what changed since version X. --- # Release Notes When invoked: 1. Run `git log --oneline <last-tag>..HEAD`. 2. Group commits by type (feat, fix, perf, docs). 3. Translate each commit subject into user-facing language. 4. Output as Markdown headed by the new version. 5. Suggest a one-line summary at the top. Output format example: see ./examples/sample-notes.md.A simple SKILL.md. The description is what makes the model find it. The body is what makes the model follow it.

When skills are the right packaging

A workflow you want to share with teammates without making them remember a custom command
A procedure that has multiple steps and benefits from explicit instructions
A task that has a 'usually trips up the model' kind of nuance worth encoding
A workflow that uses a specific template, example, or sub-script
Knowledge that should travel with the project and version-control with it

Skills you should expect to use

Type	Example skills
Personal	Daily standup writeup, weekly review
Team / project	Release notes, deploy checklist, code review rubric
Plugin	Vercel deploy, security review, Anthropic SDK migration

The big idea: skills are how teams encode 'the way we do X' into something the model can find and follow. Spend on the description; reuse the result.

End-of-lesson check

8 questions · take it digitally for instant feedback at tendril.neural-forge.io/learn/quiz/end-claude-code-skills-creators

What is the main idea of "Skills: Bundled Procedural Knowledge"?
1. Skills are reusable bundles of instructions plus optional scripts and assets.
2. Use AI as the final authority for the whole decision
3. Avoid checking the answer once it sounds polished
4. Focus only on speed instead of judgment
Which concept is most central to "Skills: Bundled Procedural Knowledge"?
1. procedural knowledge
2. skill
3. skill bundle
4. discovery
Which use of AI fits this topic best?
1. Let the AI decide what matters without your review
2. Use the answer before checking whether it fits the situation
3. User-scoped: in `~/.claude/skills/<skill-name>/` — available across all projects
4. Treat the AI output as automatically correct
What should a careful learner remember about "Skill vs slash command vs MCP"?
1. Use AI to draft or organize ideas about skill, then verify before acting.
2. Skip the context so the tool can guess faster
3. Treat the output as private even after sharing it online
4. Use the answer without checking the source
You want to use AI after this lesson. What is the safest next step?
1. Act immediately because the AI answer is written clearly
2. Use AI for drafting and comparison, but verify before publishing or relying on it.
3. Hide uncertainty so the final answer looks cleaner
4. Use private or sensitive details before checking permission
How should AI output about skill be treated?
1. As proof that no other source is needed
2. As a replacement for context, consent, or expert review
3. As a draft or helper output that still needs human judgment and verification
4. As something that becomes correct when it sounds confident
Name one way to verify an AI answer about skill.
Which action would help you apply "Skills: Bundled Procedural Knowledge" responsibly?
1. Use the tool to avoid thinking through the tradeoff
2. Keep going even if the output conflicts with a trusted source
3. Treat the AI output as automatically correct
4. Project-scoped: in `.claude/skills/<skill-name>/` inside a repo — versioned and shared with the team

← Back to interactive lesson

Tendril · Creators · Tools Literacy