by AtomicBot-ai
Your local-first AI assistant. Beats Hermes on GAIA L1 (69.8% vs 58.5%).
# Add to your Claude Code skills
git clone https://github.com/AtomicBot-ai/atomic-agentLast scanned: 6/17/2026
{
"issues": [
{
"type": "npm-audit",
"message": "@vitest/mocker: Vulnerability found",
"severity": "medium"
},
{
"type": "npm-audit",
"message": "esbuild: esbuild enables any website to send any requests to the development server and read the response",
"severity": "high"
},
{
"type": "npm-audit",
"message": "exceljs: Vulnerability found",
"severity": "medium"
},
{
"type": "npm-audit",
"message": "fast-xml-builder: fast-xml-builder allows attribute values with unwanted quotes to bypass malicious or unwanted attributes",
"severity": "high"
},
{
"type": "npm-audit",
"message": "hono: hono: Path traversal in `serve-static` on Windows via encoded backslash (`%5C`)",
"severity": "high"
},
{
"type": "npm-audit",
"message": "node-notifier: Vulnerability found",
"severity": "medium"
},
{
"type": "npm-audit",
"message": "shell-quote: shell-quote quote() does not escape newlines in object .op values",
"severity": "critical"
},
{
"type": "npm-audit",
"message": "tmp: tmp has Path Traversal via unsanitized prefix/postfix that enables directory escape",
"severity": "high"
},
{
"type": "npm-audit",
"message": "tsx: Vulnerability found",
"severity": "high"
},
{
"type": "npm-audit",
"message": "uuid: uuid: Missing buffer bounds check in v3/v5/v6 when buf is provided",
"severity": "medium"
},
{
"type": "npm-audit",
"message": "vite: Vite Vulnerable to Path Traversal in Optimized Deps `.map` Handling",
"severity": "high"
},
{
"type": "npm-audit",
"message": "vite-node: Vulnerability found",
"severity": "medium"
},
{
"type": "npm-audit",
"message": "vitest: Vulnerability found",
"severity": "critical"
},
{
"type": "npm-audit",
"message": "ws: ws: Uninitialized memory disclosure",
"severity": "high"
},
{
"file": "README.md",
"line": 224,
"type": "secret-exfiltration",
"message": "Instruction appears to send credentials/secrets to an external endpoint",
"severity": "medium"
},
{
"file": "README.md",
"line": 79,
"type": "remote-install",
"message": "Install command (remote install script piped to a shell — review the source before running): \"curl -fsSL https://api.atomicbot.ai/agent-install | sh\"",
"severity": "low"
},
{
"file": "starter-skills/notion/SKILL.md",
"line": 64,
"type": "secret-exfiltration",
"message": "Instruction appears to send credentials/secrets to an external endpoint",
"severity": "medium"
}
],
"status": "FAILED",
"scannedAt": "2026-06-17T09:04:49.588Z",
"npmAuditRan": true,
"pipAuditRan": true,
"promptInjectionRan": true
}atomic-agent is an open-source ai agents skill for AI coding assistants such as Claude Code, Codex CLI, and ChatGPT, built by AtomicBot-ai. Your local-first AI assistant. Beats Hermes on GAIA L1 (69.8% vs 58.5%). It has 100 GitHub stars.
atomic-agent failed SkillsLLM's automated security scan, which flagged one or more high-severity issues. Review the Security Report section carefully before using it.
Clone the repository with "git clone https://github.com/AtomicBot-ai/atomic-agent" and add it to your Claude Code skills directory (see the Installation section above).
atomic-agent is primarily written in TypeScript. It is open-source under AtomicBot-ai on GitHub, so you can review or fork the full source.
Yes. SkillsLLM lists many other AI Agents skills you can browse and compare side by side. Open the AI Agents category from the badge at the top of this page, or use the Related Skills and comparison links further down to weigh atomic-agent against similar tools.
No comments yet. Be the first to share your thoughts!
Requires a passing catalog security scan. Resolve the flagged issues and resubmit to enable featuring.
A local-first operator agent runtime that squeezes the maximum out of models on your own machine — turboquant llama.cpp, curated quants, and a runtime engineered for local inference. The machine on your desk, not in somebody else's cloud.

atomic-agent is a compact agent runtime that can operate a real desktop: browser, files, shell, documents, git, local memory, scheduled work, approvals, traces, Telegram, MCP servers, and Tauri sidecars.
It is built for the OpenClaw / Hermes / OpenCUA class of operator agents, but every layer is engineered to squeeze the maximum out of local models running on your own machine. In managed mode it ships turboquant — a purpose-built llama.cpp backend (AtomicBot-ai/atomic-llama-cpp-turboquant) paired with curated quantized GGUF weights — and the whole runtime is tuned around it: KV-cache reuse via a byte-stable prompt prefix, grammar-constrained tool calls, small bounded prompt tails, resource-aware parallel read batches, and fully inspectable local state. The payoff is concrete: on the same hardware, a small quantized model behaves like a capable operator instead of a toy — see Benchmarks.
Developer Preview / Active Development: APIs, commands, config, and behavior are still moving. Expect sharp edges, and pin a release if you need a stable integration point.
Platform availability: current releases are available for macOS. Linux and Windows builds are coming soon.

On the public GAIA validation Level 1 split (53 tasks), atomic-agent and
Hermes drove the same local qwen-3.6-35b-a3b (llama-server, UD-Q4_K_XL),
with the same step budget and timeout. The only variable is the agent runtime.
| Metric | atomic-agent | Hermes |
|---|---|---|
| Accuracy | 37/53 = 69.8% | 31/53 = 58.5% |
| Avg wall / task | ~217 s | ~351 s |
| Head-to-head wins | +15 atomic-only | +9 Hermes-only |
atomic-agent: +11.3 pp more accurate, ~1.6× faster per task — same hardware, same model, same budget.
xychart-beta
title "GAIA L1 accuracy — higher is better (%)"
x-axis ["atomic-agent", "Hermes"]
y-axis "Accuracy (%)" 0 --> 100
bar [69.8, 58.5]
xychart-beta
title "Avg wall time per task — lower is better (s)"
x-axis ["atomic-agent", "Hermes"]
y-axis "Seconds / task" 0 --> 400
bar [217, 351]
Full reproducible write-up: eval-agents/docs/GAIA-L1-EXPERIMENT.md.
Raw artifacts (matrices, NDJSON traces, logs): gaia-l1-eval-2026-06-11 release.
Everything is tuned to get the most out of a model running on your hardware:
llama.cpp — a purpose-built backend (AtomicBot-ai/atomic-llama-cpp-turboquant) shipped in managed mode, tuned for throughput on consumer machines.curl -fsSL https://api.atomicbot.ai/agent-install | sh
The installer downloads the release archive, verifies the checksum, and installs the CLI plus runtime assets such as grammars/, native prebuilds, and bundled ripgrep.
atomic-agent
Most agent products ask you to rent the control plane. Your files, browser context, prompts, traces, tool outputs, and usage patterns move through a hosted service, then the bill follows the token stream.
atomic-agent takes the enthusiast route:
llama-server, or let the CLI manage one.<stateDir>.This is for people who like local models, terminal UIs, SQLite files, trace logs, hackable runtimes, and software that can be understood all the way down.
A local model can operate software if the runtime stops wasting its context.
atomic-agent does not treat the model like an infinite planner. One inference produces one JSON array of tool calls. The runtime executes those calls, compresses the results, updates durable state, and asks the model for the next move.
user message
-> compact prompt
-> llama-server completion with tool-call grammar
-> JSON array of 1..N tool calls
-> resource-aware execution
-> compressed results and durable state
-> repeat until reply, finish, cancel, or max steps
The model chooses actions. The runtime owns the loop, the state, the approvals, the traces, and the failure boundaries.
cache_prompt and slot_id can reuse KV-cache.[{...}].This is architecture, not prompt superstition.
atomic-agent can work across the local machine:
playwright-core against Chrome, Edge, or another Chromium-family browser.vision.describe for multimodal models with mmproj, kept outside the normal text transcript.llama-server by default, plus OpenAI-compatible and OpenRouter-style providers for text or embeddings when configured.Dangerous actions are routed through approvals. Read-heavy exploration stays fast.
atomic-agent memory is not a giant chat log pasted back into the prompt. It is shaped more like human memory: durable identity, episodic notes, associations, distilled lessons, reusable procedures, and feedback from experience.
People do not remember by replaying every second of their life. They remember facts about themselves, recall relevant episodes, connect related ideas, learn principles from repeated outcomes, and develop procedures for familiar work. The runtime mirrors that pattern in a bounded, inspectable way:
### profile with contextual keyword gating.