Skills

Skills are reusable prompt packages that provide an agent with domain knowledge, procedures, and scripts. VeADK 1.0.0 first loads each skill’s name and description, then reads its full content when the skill is needed to limit context usage. VeADK 1.0.0 supports three execution modes:

Mode	Description
`local`	Loads skills from a local directory and executes them locally.
`skills_sandbox`	Loads skills from a cloud Skill Space and executes them in a sandbox through `execute_skills`.
`aio_sandbox`	Executes skills in an AgentKit All-in-one tool runtime.

If skills_mode is unset, VeADK selects a mode from the runtime environment: local runs use local; an AgentKit tool runtime uses skills_sandbox or aio_sandbox according to its tool type.

Skill directory

A local path in skills must point to a skills root. Each child directory represents one skill and contains SKILL.md:

skills/
└── kb-skill/
    ├── SKILL.md
    ├── references/
    ├── assets/
    └── scripts/

The SKILL.md frontmatter must define name and description, and name should match the directory name:

skills/kb-skill/SKILL.md

---
name: kb-skill
description: Query the knowledge base and compose an answer.
---

Follow these steps to query the knowledge base and compose an answer...

Use local skills

from veadk import Agent

agent = Agent(
    skills=["/abs/path/to/skills"],
    skills_mode="local",
)

Local mode adds skill metadata to the system prompt and provides the supporting tools needed to read, edit, and execute skill resources. Skill scripts can access local files and the shell environment, so load only trusted skill directories.

Use a cloud Skill Space

Pass the Skill Space identifier through skills, select skills_sandbox, and add execute_skills:

import os

from veadk import Agent
from veadk.tools.builtin_tools.execute_skills import execute_skills

agent = Agent(
    skills=[os.environ["AGENTKIT_SKILL_SPACE_ID"]],
    skills_mode="skills_sandbox",
    tools=[execute_skills],
)

Configure credentials that can access the Skill Space and AgentKit tool runtime before running the agent.

Enable a skill checklist

A skill definition can include a checklist. With enable_skills_checklist=True, VeADK initializes its items when execution begins and prompts the model to update each item through update_check_list:

from veadk import Agent

agent = Agent(
    skills=["/abs/path/to/skills"],
    skills_mode="local",
    enable_skills_checklist=True,
)

Load skills dynamically

With enable_dynamic_load_skills=True, VeADK checks and reloads skill configuration while the agent is running:

from veadk import Agent

agent = Agent(
    skills=["/abs/path/to/skills"],
    skills_mode="local",
    enable_dynamic_load_skills=True,
)

Parameters

Parameter	Type	Default	Description
`skills`	`list[str]`	`[]`	Local skills-root paths or cloud Skill Space identifiers.
`skills_mode`	`"local" \| "skills_sandbox" \| "aio_sandbox" \| None`	`None`	Execution mode; when `None`, VeADK selects it from the runtime environment.
`enable_skills_checklist`	`bool`	`False`	Tracks each item for skills that define a checklist.
`enable_dynamic_load_skills`	`bool`	`False`	Checks and reloads skills while the agent is running.

Overview

Get started

Core components

Observability

Extensions

Frontend

Security

Skill directory

Use local skills

Use a cloud Skill Space

Enable a skill checklist

Load skills dynamically

Parameters

​Skill directory

​Use local skills

​Use a cloud Skill Space

​Enable a skill checklist

​Load skills dynamically

​Parameters

Skill directory

Use local skills

Use a cloud Skill Space

Enable a skill checklist

Load skills dynamically

Parameters