"Use Atomic Agents to build schema-driven agents, integrate tools
Scanned 9/8/2026
Install to Claude Code
npx -y skills add VectorSpaceLab/AREX-Skill --skill atomic-agents --agent claude-codeInstalls into .claude/skills of the current project.
Are you the author of Atomic Agents?
Add the live security badge to your README — it updates automatically with every re-scan.
[](https://www.skillsdirectory.com/skills/vectorspacelab-atomic-agents)More formats (shields.io, HTML) on the badges page.
---
name: atomic-agents
description: "Use Atomic Agents to build schema-driven agents, integrate tools
and MCP connectors, adapt example applications, and maintain the monorepo
safely."
disable-model-invocation: true
metadata:
disco-role: operating
license: MIT
---
# Atomic Agents
Use this skill when a task mentions Atomic Agents, `atomic-agents`, `atomic-assembler`, Atomic Forge, MCP connectors, the example projects, or editing this repository.
## Start here
1. Read `references/package-summary.md` for the monorepo map, package names, and core terminology.
2. Read `references/repo-provenance.md` before deciding whether this skill matches the current checkout or needs refresh.
3. Read `references/troubleshooting.md` for cross-cutting install, import, CLI, and version-mismatch issues.
4. Use the route map below. Each subskill owns deeper references and any helper scripts.
## Install and smoke check
- Public use: `pip install atomic-agents`
- Checkout work: `uv sync` or `pip install -e .`
- Smoke check: `python -c "import atomic_agents, atomic_assembler; print(atomic_agents.__version__)"`
- CLI check: `atomic --help`
## Route map
| User task | Use subskill | Why |
| --- | --- | --- |
| Build, run, stream, or debug a typed `AtomicAgent` with schemas, history, prompts, hooks, token counting, or multimodal content | `sub-skills/agent-core/SKILL.md` | Owns the core agent API, memory, context providers, hooks, and the fundamental schema contract. |
| Define tools/resources/prompts, reason about Atomic Forge tool families, use the `atomic` CLI, or adapt tool-selection patterns | `sub-skills/tooling-and-forge/SKILL.md` | Owns `BaseTool`, `BaseResource`, `BasePrompt`, Atomic Forge, Atomic Assembler, and tool orchestration surfaces. |
| Connect to MCP servers or troubleshoot dynamic MCP-generated tools, resources, and prompts | `sub-skills/mcp-integrations/SKILL.md` | Owns the MCP transport, schema transformation, and generated tool/resource/prompt factory surface. |
| Adapt a concrete example project such as quickstart, multimodal, RAG, orchestration, memory, hooks, YouTube, DSPy, FastAPI, progressive disclosure, or `mcp-agent` | `sub-skills/example-workflows/SKILL.md` | Owns the example catalog, dependencies, and recipe-level guidance. |
| Modify this checkout, run focused tests, update docs, or reason about release/CI behavior | `sub-skills/repo-development/SKILL.md` | Owns monorepo development commands, packaging conventions, docs checks, and maintainer workflow guidance. |
## Shared references and helpers
- `references/repo-routing-metadata.json` records the router metadata consumed during import into the repo-skills router.
- `scripts/check_atomic_agents_env.py` performs a safe import/version/CLI smoke check.
- Read each subskill's `references/` and `scripts/` for workflow-specific guidance.
Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.
No comments yet. Be the first to comment!