jeevesagent.skills.tools

Tool factories that surface skills to the agent’s model.

We inject ONE tool — load_skill(name) — into the agent’s tool host whenever a non-empty SkillRegistry is configured. The tool’s input schema enumerates the registered skill names as an enum so:

  • Strict-schema providers (Anthropic / OpenAI strict mode) reject hallucinated skill names at the API boundary

  • The model sees every available skill name in the schema docs

  • Typos return a tool error with the valid set listed

The tool’s description also lists every skill with its short description, giving the model the full catalog at metadata cost without loading any bodies.

When a skill ships pending Tools (Mode B from tools.py or Mode C from frontmatter tools: manifest), load_skill ALSO registers those Tools with the agent’s tool host on the first call. The model sees the new tools in its toolset on the next turn.

Functions

make_load_skill_tool(→ jeevesagent.tools.registry.Tool)

Build the load_skill tool for a given registry.

Module Contents

jeevesagent.skills.tools.make_load_skill_tool(registry: jeevesagent.skills.registry.SkillRegistry, *, host: jeevesagent.core.protocols.ToolHost | None = None, tool_name: str = 'load_skill') → jeevesagent.tools.registry.Tool[source]

Build the load_skill tool for a given registry.

When host is provided, the tool will register a skill’s pending Tools (from Mode B / Mode C) with the host on first load — making them callable on subsequent turns. Without a host, load_skill only returns the body (skill brings no tools, or the framework integration handles registration elsewhere).