📊 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 organizing AI capabilities as ‘Skills’—folders containing instructions, scripts, and reference materials—improves consistency, onboarding, and institutional knowledge. This approach moves away from simple prompts toward durable, reusable assets that enhance organizational AI use.
Anthropic has introduced a new approach to managing AI capabilities by treating ‘Skills’ as folders—comprehensive containers of instructions, scripts, and reference materials—rather than just prompts. This shift aims to improve organizational consistency, onboarding, and knowledge retention, according to a detailed internal write-up from a Claude Code engineer. The method has been tested across hundreds of Skills within Anthropic’s engineering organization, revealing significant advantages over traditional prompt-based methods.
Most teams currently use AI coding agents by repeatedly retyping instructions, which leads to inconsistency and inefficient onboarding. Anthropic’s internal documentation describes a different approach: packaging knowledge into ‘Skills,’ which are folders containing instructions, reference documents, scripts, templates, data, and configuration. These Skills are discoverable by agents, which read and execute their contents, creating a durable, reusable asset for organizational processes.
Anthropic emphasizes that a Skill is not merely a prompt or markdown note but a container that encapsulates how a task is performed, including tribal knowledge and guardrails. This structure ensures consistent output regardless of who runs the agent, compresses onboarding by automating knowledge application, and improves over time as Skills are refined through iterative use. The company reports that its best Skills originated from small, carefully crafted starting points and improved with each edge case encountered.
Furthermore, Anthropic identified nine categories of Skills, ranging from library references and product verification to infrastructure operations. The most valuable category, according to the company, is verification—Skills that check and validate outputs—since they significantly enhance output quality and safety. The approach encourages organizations to build Skills that capture non-obvious, specific knowledge, including ‘Gotchas’—traps or pitfalls learned from experience—that prevent errors and ensure robustness.
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.
“A Skill is just a clever markdown prompt you save in a file.”
A folder the agent can discover, read & run — instructions, scripts, references, templates, config & on-demand hooks.
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.
Implications for Organizational AI Deployment
This approach represents a shift from ad-hoc prompting to structured, reusable assets that embed institutional knowledge into AI workflows. For organizations, adopting Skills as folders can lead to more consistent AI outputs, easier onboarding, and a scalable way to improve and maintain AI capabilities over time. It also encourages treating AI instructions as an asset, akin to code or documentation, rather than ephemeral prompts, potentially transforming how companies integrate AI into their operations.
AI knowledge management software
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
From Prompting to Asset-Based AI Management
Traditionally, teams rely on prompts—simple text instructions—to guide AI behavior, which can vary in effectiveness and require re-creation for each task. Anthropic’s internal experiments with hundreds of Skills demonstrate a different paradigm: organizing knowledge into structured containers that can be reused, refined, and shared across teams. This aligns with broader trends in AI development, emphasizing reliability, repeatability, and institutional memory. The concept builds on existing practices of scripting and automation but elevates them into a standardized framework for AI capabilities.
Prior to this, most organizations lacked a systematic way to codify and share AI knowledge, leading to inconsistencies and onboarding challenges. Anthropic’s findings suggest that investing time in developing comprehensive Skills can yield long-term benefits, with each iteration improving the organization’s AI performance.
“Treating Skills as folders containing instructions, scripts, and reference documents fundamentally changes how organizations embed AI capabilities into their workflows.”
— Thorsten Meyer, AI researcher
AI scripting and instruction containers
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Unanswered Questions About Skill Implementation
While Anthropic’s internal experience shows promising results, it is not yet clear how broadly this approach can be adopted across different organizations or AI systems. Details about how Skills are maintained, versioned, and scaled in larger, more complex environments remain to be seen. Additionally, the long-term impact on AI safety, performance, and cost-efficiency requires further validation in diverse operational contexts.
AI verification tools
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Next Steps for Broader Adoption and Validation
Organizations interested in this approach should evaluate how to structure their own Skills libraries, starting with key operational categories like verification and automation. Further research and case studies are expected to emerge, exploring how Skills can be integrated into existing AI workflows at scale. Anthropic likely will continue refining its methodology and sharing insights on best practices for building and maintaining Skills assets.
AI organizational knowledge platform
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Key Questions
How is a Skill different from a prompt?
A Skill is a structured folder containing instructions, scripts, and reference materials, serving as a reusable asset. In contrast, a prompt is a simple text instruction that guides AI behavior temporarily.
What benefits does organizing Skills as folders provide?
This organization improves consistency, simplifies onboarding, and allows continuous refinement, making AI capabilities more reliable and scalable within organizations.
Can this approach be applied outside of Anthropic?
While Anthropic’s internal results are promising, broader application depends on organizational capacity to develop and maintain structured Skills libraries. Further validation is needed for general use.
What are ‘Gotchas’ in Skills development?
‘Gotchas’ are traps or pitfalls, learned from experience, that prevent errors—such as misaligned data or silent failures—and are documented within Skills to improve robustness.
Will this approach reduce the need for prompt engineering?
Yes, by encapsulating knowledge into Skills, organizations can reduce ad-hoc prompt creation, leading to more stable and predictable AI outputs.
Source: ThorstenMeyerAI.com
