The freeact-skills project provides a curated set of predefined skill modules for the freeact agent system. These modules must be pre-installed on ipybox to be accessible by freeact agents.
Currently in its early development phase, freeact-skills offers a set of core skills and specialized skills. To use a skill in a freeact agent application, reference it by its module name (see freeact tutorials for details). For example, the zotero skill can be referenced as freeact_skills.zotero.api.
Skill module sources are retrieved from ipybox and incorporated into the agent's model context. By convention, freeact-skills packages are split into an api module containing the skill interface, including a detailed description and optional usage examples, and one or more implementation modules containing the underlying logic.
The freeact-skills project provides a curated set of predefined skill modules for the freeact agent system. These modules must be pre-installed on ipybox to be accessible by freeact agents.
Currently in its early development phase, freeact-skills offers a set of core skills and specialized skills. To use a skill in a freeact agent application, reference it by its module name (see freeact tutorials for details). For example, the zotero skill can be referenced as freeact_skills.zotero.api.
Skill module sources are retrieved from ipybox and incorporated into the agent's model context. By convention, freeact-skills packages are split into an api module containing the skill interface, including a detailed description and optional usage examples, and one or more implementation modules containing the underlying logic.
This convention ensures that freeact agents only receive relevant information from the api module while internal implementation details remain hidden. However, it is merely a guideline; you are free to present any Python code as a skill module, depending on your context requirements and the level of detail you want the agent to process.
freeact_skills.search.google.stream.api
Provides generative Google search via the Gemini 2 API. Search output is streamed, enabling clients to display results as they are generated.
freeact_skills.search.google.api
Provides generative Google search via the Gemini 2 API. Search output is returned as a string for direct processing by agents in code actions. Maintained for experimentation with Gemini 2 based freeact agents.
Both Google search skills require a GOOGLE_API_KEY or GEMINI_API_KEY, obtainable from Google AI Studio.
freeact_skills.search.perplexity.api
Provides generative Perplexity search via the Perplexity API. Search output is streamed for real-time display and includes indexed source links in results. Requires a PERPLEXITY_API_KEY, obtainable from the Perplexity API Settings page.
freeact_skills.zotero.api
Provides access to a Zotero group library via the Zotero API. Requires a ZOTERO_API_KEY which can be obtained from the Zotero Security Settings page. A ZOTERO_GROUP_ID environment variable with a group library ID must also be defined. Here's what a freeact agent says about this skill module when asked to describe it:
The Zotero module provides a way to interact with Zotero group libraries programmatically. Here's what it can do in simple terms:
Basic Library Access
Document Management
Collection Organization
Attachments
The module is primarily designed for reading and organizing academic references in a Zotero group library, with some ability to add new content (like links). It maintains a local copy of the library that can be synchronized with Zotero's servers to get the latest updates.
freeact_skills.reader.api
Provides access to Readwise Reader via the Readwise Reader API. Requires a READWISE_API_KEY which can be obtained from the here. Here's what a freeact agent says about this skill module when asked to describe it:
The reader module provides the following capabilities:
Save documents to Readwise Reader in two ways:
save_document_url(). You can optionally specify:save_document_html() with similar options as above, plus requiring a titleList and retrieve documents from Readwise Reader using list_documents():
The module provides a factory function create_readwise_reader() to instantiate the client:
This module essentially allows you to programmatically save documents to Readwise Reader and retrieve information about your saved documents, enabling automation of document management and reading workflows.
freeact_skills.editor
Internal system component (not for direct application use). Enables agents to view, create, edit, insert, and undo file edits, supporting interactive skill module development. Based on Anthropic's original implementation.
freeact-skills is in early development. Skill interfaces and implementations may change frequently, relying on the adaptability of freeact agents to handle such changes gracefully.
For use with freeact agents, freeact-skills needs to be pre-installed on ipybox Docker images. First, install ipybox:
pip install ipybox\nCreate a dependencies.txt file with one of the following configurations:
# Install all available skills\nfreeact-skills = {version = \"*\", extras = [\"all\"]}\nThe extras=[\"all\"] option includes all available skills: [\"reader\", \"search-google\", \"search-perplexity\", \"zotero\"]. Alternatively, you can install specific skills:
# Install selected skills only\nfreeact-skills = {version = \"*\", extras = [\"search-google\", \"zotero\"]}\nNote
dependencies.txt must follow the Poetry dependency specification format.
Build the ipybox Docker image with your selected skills:
python -m ipybox build -t your-image-tag -d dependencies.txt\nReplace your-image-tag with your preferred image name.
For local development or direct usage, install with pip:
pip install freeact-skills[all]\nThis is equivalent to:
pip install freeact-skills[reader,search-google,search-perplexity,zotero]\nTo install specific skills only:
pip install freeact-skills[search-google,zotero]\nThe freeact-skills project provides a curated set of predefined skill modules for the freeact agent system. These modules must be pre-installed on ipybox to be accessible by freeact agents.
Currently in its early development phase, freeact-skills offers a set of core skills and specialized skills. To use a skill in a freeact agent application, reference it by its module name (see freeact tutorials for details). For example, the zotero skill can be referenced as freeact_skills.zotero.api.
Skill module sources are retrieved from ipybox and incorporated into the agent's model context. By convention, freeact-skills packages are split into an api module containing the skill interface, including a detailed description and optional usage examples, and one or more implementation modules containing the underlying logic.
This convention ensures that freeact agents only receive relevant information from the api module while internal implementation details remain hidden. However, it is merely a guideline; you are free to present any Python code as a skill module, depending on your context requirements and the level of detail you want the agent to process.
freeact_skills.search.google.stream.api
Provides generative Google search via the Gemini 2 API. Search output is streamed, enabling clients to display results as they are generated.
freeact_skills.search.google.api
Provides generative Google search via the Gemini 2 API. Search output is returned as a string for direct processing by agents in code actions. Maintained for experimentation with Gemini 2 based freeact agents.
Both Google search skills require a GOOGLE_API_KEY or GEMINI_API_KEY, obtainable from Google AI Studio.
freeact_skills.search.perplexity.api
Provides generative Perplexity search via the Perplexity API. Search output is streamed for real-time display and includes indexed source links in results. Requires a PERPLEXITY_API_KEY, obtainable from the Perplexity API Settings page.
freeact_skills.zotero.api
Provides access to a Zotero group library via the Zotero API. Requires a ZOTERO_API_KEY which can be obtained from the Zotero Security Settings page. A ZOTERO_GROUP_ID environment variable with a group library ID must also be defined. Here's what a freeact agent says about this skill module when asked to describe it:
The Zotero module provides a way to interact with Zotero group libraries programmatically. Here's what it can do in simple terms:
Basic Library Access
Document Management
Collection Organization
Attachments
The module is primarily designed for reading and organizing academic references in a Zotero group library, with some ability to add new content (like links). It maintains a local copy of the library that can be synchronized with Zotero's servers to get the latest updates.
freeact_skills.reader.api
Provides access to Readwise Reader via the Readwise Reader API. Requires a READWISE_API_KEY which can be obtained from the here. Here's what a freeact agent says about this skill module when asked to describe it:
The reader module provides the following capabilities:
Save documents to Readwise Reader in two ways:
save_document_url(). You can optionally specify:save_document_html() with similar options as above, plus requiring a titleList and retrieve documents from Readwise Reader using list_documents():
The module provides a factory function create_readwise_reader() to instantiate the client:
This module essentially allows you to programmatically save documents to Readwise Reader and retrieve information about your saved documents, enabling automation of document management and reading workflows.
freeact_skills.editor
Internal system component (not for direct application use). Enables agents to view, create, edit, insert, and undo file edits, supporting interactive skill module development. Based on Anthropic's original implementation.
freeact-skills is in early development. Skill interfaces and implementations may change frequently, relying on the adaptability of freeact agents to handle such changes gracefully.
For use with freeact agents, freeact-skills needs to be pre-installed on ipybox Docker images. First, install ipybox:
pip install ipybox\nCreate a dependencies.txt file with one of the following configurations:
# Install all available skills\nfreeact-skills = {version = \"*\", extras = [\"all\"]}\nThe extras=[\"all\"] option includes all available skills: [\"reader\", \"search-google\", \"search-perplexity\", \"zotero\"]. Alternatively, you can install specific skills:
# Install selected skills only\nfreeact-skills = {version = \"*\", extras = [\"search-google\", \"zotero\"]}\nNote
dependencies.txt must follow the Poetry dependency specification format.
Build the ipybox Docker image with your selected skills:
python -m ipybox build -t your-image-tag -d dependencies.txt\nReplace your-image-tag with your preferred image name.
For local development or direct usage, install with pip:
pip install freeact-skills[all]\nThis is equivalent to:
pip install freeact-skills[reader,search-google,search-perplexity,zotero]\nTo install specific skills only:
pip install freeact-skills[search-google,zotero]\n