> For the complete documentation index, see [llms.txt](https://docs.eesel.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.eesel.ai/skills/skills.md).
# Skills
A skill is a job your teammate knows how to do end to end: a multi-step piece of work with its own process, its own output, and its own place to put that output. Testing itself against hundreds of your past tickets is a skill. Reading a month of conversations and writing up what customers keep asking about is a skill.
And a skill doesn't just report back. It can take action on what it found, so the work finishes rather than landing in your lap.
## Quick start
{% stepper %}
{% step %}
### Chat with your agent to start a skill, or create one
Say *"run a simulation on my recent tickets"* or *"tell me what customers have been asking about"* and your teammate adds the skill it needs and gets on with the job. You don't have to know what the skill is called. Prefer to browse? Skills live on your agent's **Skills** page in the left sidebar, added per agent so each agent carries the ones that fit its job.
{% endstep %}
{% step %}
### Clarify the details and connect what it needs
The skill asks what it needs to know, and you answer in plain language: which cases to cover, what the output should look like, any parameters you care about. If the job needs a platform you haven't connected (your helpdesk, your team chat, or an outside API through [Network Access](/apis-and-developer-resources/network-access.md)), it tells you and walks you through connecting it.
{% endstep %}
{% step %}
### Let it run
It tells you when it's done, in chat, or wherever you asked to be told. Every run also lands in **Activity** with a **View Output** badge, so nothing gets lost.
Skills don't have to run on demand, either: attach one to an [automation](/automations-and-schedules/automations-and-schedules.md) and it runs on a schedule or off an event. A report landing every Monday before your team meeting is part of how your team works.
{% endstep %}
{% step %}
### Review and act on the output
Send the result where the people who need it already are:
| Send it | Good for |
| ----------------------------- | ---------------------------------------------------------------------------------------------- |
| **A team chat channel** | Reports the whole team should see |
| **A direct message or email** | Something only one person needs |
| **A ticket** | Work that should enter your support queue |
| **Your dashboard** | Reading it yourself, in full |
| **A file in eesel** | Keeping it for later. A saved report becomes knowledge your agent can compare against next run |
Or have it **act on what it found**: write a corrected rule into its own [instructions](/instructions-and-memory/instructions-and-memory.md), refresh a metrics file, create or route tickets, or flag the person who owns the problem.
{% endstep %}
{% endstepper %}
## What ships out of the box
| Skill | What it does |
| ------------------------------- | ----------------------------------------------------------------------------------------------------------------------- |
| **Simulation** | Runs your agent against real past tickets or generated test cases, scores each answer, and suggests instruction changes |
| **Analyze and improve replies** | Looks at what your team rejected or edited, finds the pattern, and suggests fixes |
| **Review agent setup** | Audits your agent's configuration and performance, and tells you what to change |
| **Analyze ticket trends** | Recurring topics, sentiment, volume and resolution patterns across your conversations |
| **Update knowledge base** | Finds the questions your help center doesn't answer, and drafts the missing articles |
| **Write a blog post** | Researches and writes a full post with images. See [Content Writing](/use-cases/content-writing.md) |
| **Translate content** | Translates content while keeping formatting and meaning intact |
| **Draft an internal note** | Writes a reply as an internal note for your team to review |
| **Send a public reply** | Sends a reply to the customer |
| **Look up Calendly meetings** | Checks your Calendly for meetings and availability |
| **Voice call transcriber** | Transcribes call recordings |
| **Commit and push to GitHub** | Commits and pushes changes to a repository |
## Custom skills, your way
{% hint style="success" %}
**If your team does it, it can be a skill.** Connect any platform, build any report, complete any task end to end: your weekly numbers in your format, cross-system checks between your helpdesk and your billing system, competitor watch, escalation triage, or anything your own tools expose over an API. Describe the job in chat and your teammate builds the skill with you. See [Custom Skills](/skills/custom-skills.md).
{% endhint %}
## FAQ
A skill shows "Contact us" instead of an add button. Why?
Some skills need to be switched on for your workspace before you can use them. If one isn't available, [contact us](https://eesel.ai/contact) or ask in the in-app support chat under **Help and support**.
Is there a community marketplace?
The skills on your Skills page are eesel-built, installable there or through the [CLI](/apis-and-developer-resources/cli.md). A community marketplace where teams publish skills for each other is something we're working towards. For now the catalogue is ours plus whatever you build.
Can a skill work across more than one tool?
Yes. A skill can use anything your agent is connected to, in one run: read your helpdesk, check Shopify, and post the result to Slack. Custom skills can also reach your own systems through [Network Access](/apis-and-developer-resources/network-access.md).
Where do I see what a skill produced?
Every run shows up in **Activity** with its output attached, and a skill's own page lists its past runs. Outputs sent to a file live under **Files**, where your agent can read them again later.
Can a skill have its own instructions?
Yes. A skill can carry saved instructions that apply only when it runs, so a blog style guide doesn't sit in the instructions your support replies read. Your agent's core [instructions](/instructions-and-memory/instructions-and-memory.md) still apply underneath.
## Related
* [Custom Skills](/skills/custom-skills.md) to build your own
* [Automations and Schedules](/automations-and-schedules/automations-and-schedules.md) to run a skill on a schedule or a trigger
* [Instructions and Memory](/instructions-and-memory/instructions-and-memory.md) for how skills feed back into how your agent behaves
---
# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.
## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.
Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:
```
GET https://docs.eesel.ai/skills/skills.md?ask=&goal=
```
`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.
Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.