AI Skills (lexigram-ai-skills)

AI skills and tools for the Lexigram Framework — registry, executor, builtin tools, discovery

Overview

Composable, registry-based skill execution for the Lexigram AI framework. Define skills as classes or decorated functions, execute them with retry, caching, permission enforcement, and timeout — and compose them into chains, pipelines, parallel fans, and content-routers. Zero-config usage starts with sensible defaults.

Install

uv add lexigram-ai-skills

Quick Start

from lexigram import Application
from lexigram.di.module import Module, module

from lexigram.ai.skills import SkillsModule
from lexigram.ai.skills.config import SkillsConfig

@module(imports=[
    SkillsModule.configure(
        SkillsConfig(
            enable_builtin=True,
            builtin_skills=["current_datetime", "math_calculate"],
            cache_enabled=True,
            enforce_permissions=False,
        )
    )
])
class AppModule(Module):
    pass

app = Application(modules=[AppModule])
if __name__ == "__main__":
    app.run()

Configuration

Zero-config usage: Call SkillsModule.configure() with no arguments to use defaults.

Option 1 — YAML file

ai_skills:
  default_timeout_seconds: 30.0
  cache_enabled: true
  enforce_permissions: true
  enable_builtin: true
  builtin_skills:
    - "current_datetime"
    - "math_calculate"

Option 2 — Profiles + Environment Variables (recommended)

export LEX_AI_SKILLS__ENFORCE_PERMISSIONS=true
# Environment variables for each field

Option 3 — Python

from lexigram.ai.skills.config import SkillsConfig
from lexigram.ai.skills import SkillsModule

config = SkillsConfig(
    default_timeout_seconds=30.0,
    cache_enabled=True,
    enforce_permissions=True,
    enable_builtin=True,
    builtin_skills=["current_datetime", "math_calculate", "text_summarize"],
)
SkillsModule.configure(config)

Config reference

Field	Default	Env var	Description
`default_timeout_seconds`	`30.0`	`LEX_AI_SKILLS__DEFAULT_TIMEOUT_SECONDS`	Execution timeout per skill
`max_retries`	`2`	`LEX_AI_SKILLS__MAX_RETRIES`	Retry attempts on failure
`max_concurrent_executions`	`10`	`LEX_AI_SKILLS__MAX_CONCURRENT_EXECUTIONS`	Semaphore cap on concurrent executions
`cache_enabled`	`True`	`LEX_AI_SKILLS__CACHE_ENABLED`	Global result caching switch
`cache_ttl_seconds`	`3600`	`LEX_AI_SKILLS__CACHE_TTL_SECONDS`	Default TTL for cached results
`enforce_permissions`	`True`	`LEX_AI_SKILLS__ENFORCE_PERMISSIONS`	Enable permission checks
`auto_discover`	`False`	`LEX_AI_SKILLS__AUTO_DISCOVER`	Scan `scan_packages` on boot
`enable_builtin`	`True`	`LEX_AI_SKILLS__ENABLE_BUILTIN`	Register built-in skills on boot
`builtin_skills`	(list)	`LEX_AI_SKILLS__BUILTIN_SKILLS`	Which built-in skills to register
`enable_skill_sources`	`True`	`LEX_AI_SKILLS__ENABLE_SKILL_SOURCES`	Scan external SKILL.md sources on boot

Module Factory Methods

Method	Description
`SkillsModule.configure(config)`	Configure with explicit config
`SkillsModule.stub(config)`	Minimal config for testing

Key Features

Class-based skills: AbstractSkill base class with SkillDefinition
Function-based skills: @skill decorator for registering functions
Skill executor: Retry, caching, permissions, and timeout enforcement
Built-in skills: DateTime, Math, HTTPRequest, WebSearch, FileRead, FileWrite, DatabaseQuery, CodeExecute
Composition: SkillChain, ParallelSkills, SkillPipeline, SkillRouter
MCP bridge: MCPSkillBridge imports MCP tools as skills and exports skills as MCP tools
Registry: Named lookup and registration via SkillRegistry

Testing

async with Application.boot(modules=[SkillsModule.stub()]) as app:
    # your test code
    ...

Key Source Files

File	What it contains
`src/lexigram/ai/skills/module.py`	`SkillsModule.configure()`, `.stub()`
`src/lexigram/ai/skills/config.py`	`SkillsConfig`
`src/lexigram/ai/skills/base/core.py`	`AbstractSkill`, `FunctionSkill`
`src/lexigram/ai/skills/decorators/core.py`	`@skill`, `@skill_param`
`src/lexigram/ai/skills/executor/core.py`	`SkillExecutor`
`src/lexigram/ai/skills/registry/core.py`	`SkillRegistry`
`src/lexigram/ai/skills/composition/`	chain, parallel, pipeline, router
`src/lexigram/ai/skills/discovery/mcp_bridge.py`	`MCPSkillBridge`
`src/lexigram/ai/skills/di/provider.py`	`SkillsProvider`