Agent Skills by Anthropic on Oct 16, 2025, as a approach to lengthen Claude with reusable capabilities. Inside months, the idea gained traction throughout the AI group and commenced evolving right into a broader design sample for constructing modular, transportable agent capabilities past Claude itself.
As an AI practitioner, I’ve been utilizing Claude Code for fairly a while and have seen many tutorials explaining the best way to create Agent Abilities throughout the Claude ecosystem. Nonetheless, after I tried to implement the identical idea for our enterprise solution with out counting on Claude, I rapidly bumped into a special set of design questions. What precisely defines an “agent talent”? How ought to or not it’s structured? And the way ought to expertise be triggered and orchestrated in a customized agent framework?
In Claude, a typical Agent Ability is outlined as a Markdown file containing a reputation, description, directions, and scripts. This natural-language interface works nicely in lots of conditions, particularly when ambiguity is appropriate. However in manufacturing methods, precision typically issues greater than flexibility. One recurring problem I encountered was talent triggering: typically a talent could be highly effective and well-designed, but Claude would fail to invoke it as a result of the outline was too imprecise.
For our brokers, I would like stricter ensures—expertise that set off deterministically and behave persistently each time. That requirement led me to implement expertise instantly in Python with express logic. However doing so raised an attention-grabbing architectural query: if a talent is applied purely in code, how is it completely different from a instrument, a function, or simply one other perform? Even when these expertise are reusable and invoked on demand, what really makes them agent expertise?
This text explores these questions and shares a sensible perspective on designing and implementing agent expertise for customized AI brokers with out counting on Claude.
Agent expertise vs Instruments vs Options
A instrument is a primitive functionality, or a single motion the agent can take. Every instrument does one particular factor. Instruments are particular person devices, like hammers, saws, and drills.
- Instance: Claude’s instruments embody issues like
bash_tool(run a command),web_search(search the web),view(learn a file),str_replace(edit a file),web_fetch(get a webpage),places_search,weather_fetchand so forth.
A talent is a set of directions for the best way to orchestrate a number of instruments to perform a fancy job nicely. A talent doesn’t give me new capabilities. It offers brokers experience in combining current instruments successfully. Abilities are like an in depth recipe that tells the brokers which devices to make use of, in what order, and what errors to keep away from.
- Instance: The docx talent, for instance, tells Claude to make use of
bash_toolto runnpm set up docx, then write JavaScript utilizing particular patterns, then runbash_toolonce more to validate the output, then usepresent_filesto share it.
A function is a product-level idea, or one thing the person sees and may toggle on or off. A function is enabled by giving the agent entry to sure instruments and expertise.
- Examples: “Code Execution and File Creation,” “Internet Search,” and “Artifacts” are options. So “file creation” as a function is powered by the bash instrument, the file creation instrument, numerous doc expertise, and the present_files instrument, all working collectively.
Abilities are what bridge the hole between uncooked instrument entry and high-quality output. With out the docx talent, Claude may nonetheless technically create a Phrase doc, but it surely may miss issues like “all the time set web page measurement explicitly as a result of docx-js defaults to A4” or “by no means use unicode bullets.”
Do agent expertise must be within the format of markdown information?
Not essentially. The idea of agent expertise is broader than the format.
A talent is essentially codified experience for the best way to accomplish a job nicely. That experience may dwell in lots of kinds:
- A markdown file with directions (typical Claude Agent Ability)
- A Python script that does the work instantly
- A config file or JSON schema
- A set of instance inputs and outputs
- A mix of all the above
In my case, the Python script appears applicable as a result of I would like deterministic and dependable execution each single time with no variation. It’s sooner, cheaper, and extra predictable for a procedural course of. The markdown instruction strategy turns into beneficial when the duty includes ambiguity or judgment. Typically, we want LLM studying directions and reasoning about what to do subsequent so as to add worth over a inflexible script.
A hybrid strategy can be frequent. I can maintain my Python code and downgrade it to a instrument implementation, however add a markdown talent file that helps the brokers perceive when to invoke my agent talent and the best way to interpret the outcomes. A typical product iteration may start with a Python implementation and progressively incorporate Markdown directions, finally forming a hybrid talent design.
Agent talent vs MCP? Agent Ability + MCP!
Our brokers already hook up with knowledge sources by MCP servers, however lots of our agent expertise additionally embody the power to learn instantly from databases. This raises a sensible architectural query: when ought to an agent use MCP, and when ought to the aptitude dwell inside a talent?
Anthropic explains the excellence clearly with a useful kitchen analogy:
MCP connects Claude to knowledge; Abilities teaches Claude what to do with that knowledge.
- MCP offers the skilled kitchen – entry to instruments, elements, and gear.
- Abilities present the recipes – directions that inform the agent the best way to use these assets to provide one thing beneficial.
In my design, I deal with MCP connections as infrastructure and expertise as orchestration on how the info is used.
MCP servers are liable for exposing exterior knowledge sources and providers. For instance, they could present structured entry to databases, logs, APIs, or inside methods. Their function is to make these assets obtainable to the agent in a standardized means.
Agent expertise, however, outline how the agent ought to use the info from MCP servers and different databases to perform a job.
Because of this, I usually implement:
- Database entry, APIs, and knowledge retrieval as instruments (or MCP capabilities)
- Determination logic and workflows as agent expertise
Abilities as “Agentic RAG”
Ideally, the agent expertise load data dynamically from instruments when they’re executed. This makes the sample similar to agentic retrieval-augmented technology (RAG).
As an alternative of preloading all context into the immediate, the talent can:
- Establish what data it wants
- Retrieve the related knowledge by a instrument or MCP server
- Course of that knowledge in keeping with its directions
- Produce the ultimate output
This strategy retains the brokers light-weight whereas nonetheless permitting expertise to entry giant or altering datasets on demand.
Conclusion
Anthropic launched an vital paradigm shift, and Claude’s implementation of Agent Abilities affords beneficial inspiration for constructing our personal agent methods. So long as our expertise seize the core goal of what a talent is supposed to do, which is encapsulating reusable on-demand capabilities for an agent, the particular format and implementation particulars can differ. In the end, the design choices round when and the best way to use expertise needs to be guided by the wants and constraints of the product we’re constructing.
Thanks for studying! I hope this has been useful to you.
