A Skill Is A Folder, Not A Prompt: What Anthropic Learned Running Hundreds Of Them

📊 Full opportunity report: A Skill Is A Folder, Not A Prompt: What Anthropic Learned Running Hundreds Of Them on ThorstenMeyerAI.com — validation score, market gap, and execution plan.

TL;DR

Anthropic has demonstrated that modeling AI skills as folders containing instructions, scripts, and assets improves consistency, onboarding, and institutional memory. This approach shifts from ad-hoc prompting to durable, reusable organizational units.

Anthropic has revealed that effective AI skills are best conceptualized as folders containing instructions, scripts, and reference materials, rather than just saved prompts. This approach, based on internal experiments, aims to create durable, reusable units that improve consistency and onboarding in AI workflows, marking a significant shift in how organizations develop and deploy AI capabilities.

In a detailed write-up from a Claude Code engineer, Anthropic explains that a Skill is not merely a prompt but a folder that can include instructions, reference documents, scripts, templates, data, configuration, and hooks. This structure allows AI agents to discover, read, and execute contained assets, making the process more reliable and maintainable.

Anthropic’s internal experiments involved running hundreds of Skills across its engineering organization, leading to a classification into nine categories, from library references to infrastructure operations. The most impactful category identified was verification Skills, which ensure output quality by catching mistakes before they reach production.

According to the company, developing Skills as containers helps standardize output, reduce onboarding time, and enable continuous improvement through iteration. The approach turns ad-hoc prompting into a formalized, institutional capability, akin to maintaining a shared asset library rather than individual notes.

At a glance
reportWhen: published recently, with ongoing implem…
The developmentAnthropic published insights from running hundreds of AI ‘Skills’ as folders, redefining how organizations build and maintain AI capabilities.
A Skill Is a Folder, Not a Prompt — Insights
AI Dispatch · Insights · 1 July 2026

A Skill is a folder, not a prompt

Anthropic published what it learned running hundreds of Skills across its own engineering org. Read as a business memo, the point is bigger than a coding trick: this is how ad-hoc prompting becomes durable institutional capability — the SOPs your agents actually follow, versioned and shared.

✕ The misconception

“A Skill is just a clever markdown prompt you save in a file.”

✓ What it actually is

A folder the agent can discover, read & run — instructions, scripts, references, templates, config & on-demand hooks.

Anatomy of a Skill — the file system is context engineering
my-skill/the unit you share & version
├─ SKILL.mdroot instructions + a description written for the model (its trigger)
├─ references/deep detail pulled in only when needed — progressive disclosure
├─ scripts/real code, so the agent composes instead of rebuilding boilerplate
├─ assets/templates & files to copy into the output
├─ config.jsonsetup the agent asks for if it’s missing (e.g. which Slack channel)
└─ hooks + memoryon-demand guardrails + an append-only log so it remembers
Why it matters: the folder itself is the knowledge base. The agent reads the root, then reaches deeper only when the task demands it — the same way you’d hand a new hire a one-pager that points to the detailed docs.
The nine types — a gap-analysis map for your own library
1Library / API reference
2Product verification ★ top impact
3Data fetching & analysis
4Business-process automation
5Code scaffolding & templates
6Code quality & review
7CI/CD & deployment
8Runbooks
9Infrastructure operations
By Anthropic’s own measurement, verification Skills — the ones that check the work — moved output quality the most. If you build one category well, build that one.
The craft — what separates a good Skill from a useless one
Gotchas = highest-signal section Describe for the model, not humans (it’s the trigger) Don’t state the obvious Ship scripts, not just prose On-demand guardrail hooks (/careful, /freeze) Let it remember (log / SQLite) Don’t railroad — leave room to adapt
The take

The knowledge of how your organization actually operates can be captured, versioned, shared & executed — and the thing capturing it is a humble folder with a script and a gotchas list inside. For the builder, that’s context engineering with real tools attached. For whoever owns the budget, it’s the difference between AI that starts from zero every morning and an asset that compounds. Caveats: best practices are still evolving, checked-in Skills cost context, and curation beats accumulation. Start with one Skill, one gotcha, and the category that catches your mistakes.

Source: “Lessons from building Claude Code: How we use skills,” Thariq Shihipar (Anthropic), Claude blog, 3 June 2026. Categories, examples & measured claims are Anthropic’s; framing is the author’s. Docs: code.claude.com/docs/en/skills.
thorstenmeyerai.com

Transforming AI Workflow Management with Folder-Based Skills

This development matters because it shifts the paradigm from ephemeral, prompt-based interactions to durable organizational assets that embed tribal knowledge, guardrails, and tools. For businesses, this means more consistent AI outputs, faster onboarding, and a scalable way to improve AI capabilities over time. It also suggests a path toward more reliable, maintainable AI systems as organizations grow more dependent on automation and AI-driven processes.

Amazon

AI development folder structure tools

As an affiliate, we earn on qualifying purchases.

As an affiliate, we earn on qualifying purchases.

From Prompt Engineering to Organizational Asset Libraries

Prior to this insight, most teams relied on prompt engineering—crafting and reusing prompts to guide AI behavior. However, this approach is fragile and hard to scale. Anthropic’s experiments with running hundreds of Skills as folders emerged from internal efforts to institutionalize AI knowledge, moving beyond ad-hoc prompts toward structured, reusable units. The concept aligns with broader trends in software engineering, where modular, versioned assets improve reliability and scalability.

This approach builds on prior work in AI prompt management but emphasizes the importance of comprehensive containers that include scripts, reference materials, and configuration, enabling more complex and dependable workflows.

“A Skill is a folder — one that can contain instructions, reference documents, runnable scripts, templates, data, configuration, and hooks that fire only while the Skill is active.”

— Thorsten Meyer, AI researcher

Unclear Aspects of Scaling Folder-Based Skills

It remains unclear how widely organizations will adopt this approach outside Anthropic and how it will scale across different industries. The specific technical challenges of managing large libraries of Skills, version control, and integration with existing systems are still being explored. Additionally, the long-term impact on AI behavior consistency and maintenance costs requires further validation.

Next Steps for Implementing Folder-Based Skills

Organizations are expected to experiment with creating their own Skills libraries based on this model, focusing on categories most relevant to their operations. Anthropic plans to continue refining its methodology, sharing best practices, and developing tooling to support large-scale deployment. Future updates may include standardized frameworks and integration guides to facilitate adoption across sectors.

Key Questions

How does a folder-based Skill differ from traditional prompt engineering?

A folder-based Skill is a structured container that includes instructions, scripts, and reference materials, making it a reusable, durable asset. In contrast, prompt engineering relies on crafting specific prompts that are often ephemeral and less maintainable.

What are the main benefits of using Skills as folders?

Using Skills as folders improves output consistency, reduces onboarding time, and enables continuous improvement through iteration. It also helps institutionalize tribal knowledge and operational guardrails.

What categories of Skills did Anthropic identify?

Anthropic classified Skills into nine categories, including library references, product verification, data analysis, business process automation, code scaffolding, quality review, deployment, runbooks, and infrastructure operations.

Are there technical challenges to adopting this approach?

Yes, managing large libraries of Skills, ensuring version control, and integrating with existing tools are potential challenges. Further research and tooling are needed to scale this approach effectively.

Will this method work for organizations outside Anthropic?

While promising, adoption depends on organizational needs, technical infrastructure, and willingness to invest in building and maintaining Skills libraries. Broader industry validation is ongoing.

Source: ThorstenMeyerAI.com

You May Also Like

A Light-Driven Chip Could Redefine Energy Use in AI Systems.

What if AI chips used light instead of electricity, drastically reducing energy consumption and transforming data processing—discover how this innovation could change everything.

The Labor Displacement Data: What Q1-Q2 2026 Actually Shows

New data from early 2026 shows significant AI-driven layoffs in tech, with impacts concentrated among entry-level and junior roles, indicating structural labor shifts.

The Local-First Agentic Operator

A new paradigm allows a single operator, using agentic AI, to build and manage multiple software products across domains, traditionally requiring teams.

Avaoroi Analysis: GitLab’s AI Investment Pays Off Big Time – Future Prospects!

On the heels of impressive revenue growth, GitLab’s AI investment raises questions about its future strategies and competitive landscape. What challenges could emerge next?