by dragonked2
Free MIT AI coding agent — no API key needed. Built-in free model, or bring Claude, GPT, Gemini, DeepSeek, Ollama +50 more. Swarm mode, 40+ tools, browser & desktop automation, built in Rust.
# Add to your Claude Code skills
git clone https://github.com/dragonked2/alphacodeLast scanned: 9/24/2026
{
"issues": [
{
"file": "README.md",
"line": 194,
"type": "remote-install",
"message": "Install command (remote install script piped to a shell — review the source before running): \"curl -fsSL https://raw.githubusercontent.com/dragonked2/alphacode/main/scripts/i\"",
"severity": "low"
},
{
"file": "src/alphacode_base/bundled_skills/bug-bounty-methodology/SKILL.md",
"line": 157,
"type": "secret-exfiltration",
"message": "Instruction appears to send credentials/secrets to an external endpoint",
"severity": "medium"
}
],
"status": "PASSED",
"scannedAt": "2026-09-24T09:01:47.029Z",
"npmAuditRan": true,
"pipAuditRan": true,
"promptInjectionRan": true
}See how alphacode compares with popular alternatives.
alphacode is an open-source ai agents skill for AI coding assistants such as Claude Code, Codex CLI, and ChatGPT, built by dragonked2. Free MIT AI coding agent — no API key needed. Built-in free model, or bring Claude, GPT, Gemini, DeepSeek, Ollama +50 more. Swarm mode, 40+ tools, browser & desktop automation, built in Rust. It has 233 GitHub stars.
Yes. alphacode passed SkillsLLM's automated security scan — a dependency vulnerability audit plus prompt-injection heuristics — with no high-severity issues. You can read the full report in the Security Report section on this page.
Clone the repository with "git clone https://github.com/dragonked2/alphacode" and add it to your Claude Code skills directory (see the Installation section above).
alphacode is primarily written in Rust. It is open-source under dragonked2 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 alphacode against similar tools.
No comments yet. Be the first to share your thoughts!
⚠️ Third-Party Software Notice
This skill is third-party open-source software developed and hosted independently on GitHub. SkillsLLM is an informational directory and does not control or maintain the underlying repository.
Any security checks, ratings, or warnings displayed by SkillsLLM are automated and limited in scope. They do not constitute a security certification or guarantee that the software is safe, error-free, or free from malicious code, vulnerabilities, compromised dependencies, or prompt-injection risks.
Review the source code, permissions, dependencies, and configuration before installing or running any third-party skill. Use is at your own risk. To the maximum extent permitted by applicable law, SkillsLLM is not liable for losses arising from third-party software.
A free, Rust-native AI coding agent for your terminal. It reads your codebase, edits files, runs commands and tests, drives a real Firefox browser, controls desktop apps, and verifies its own work.
Official website · GitHub repository
What is AlphaCode? · Quick Start · Features · Browser Agent · Desktop Control · Swarm Mode · Providers · Benchmarks · Uninstall · FAQ
AlphaCode is an open-source AI coding agent that works inside your development environment instead of just chatting about it. Describe a goal in plain English and it investigates, plans, edits files, runs commands, uses the web, operates a real browser, interacts with desktop apps, runs your tests, and reports back with evidence.
flowchart LR
A["You<br/>describe a goal"] --> B["Understand<br/>project + context"]
B --> C["Plan<br/>choose actions"]
C --> D["Act<br/>code, browser, desktop"]
D --> E["Verify<br/>tests + evidence"]
E --> F["Report<br/>what changed"]
Most AI assistants stop at You → AI → text. AlphaCode is built around a loop: understand → plan → execute → observe → verify.
Example tasks:
Find why the app crashes when users upload a PDF.
Fix the root cause, add a regression test, run the test suite, and review the final diff.
Open the site in Firefox, log in with my existing session,
find the account settings page, check that the new feature is visible,
and send me a screenshot.
Audit this project for security problems. Trace untrusted input to sensitive operations,
validate the realistic findings, and tell me which ones are actually reproducible.
Who it is for: beginners can use plain English, developers can give precise technical instructions, and power users can orchestrate multiple agents, skills and MCP servers.
macOS / Linux
curl -fsSL https://raw.githubusercontent.com/dragonked2/alphacode/main/scripts/install.sh | bash
Windows (PowerShell)
irm https://raw.githubusercontent.com/dragonked2/alphacode/main/scripts/install.ps1 | iex
Prefer to read a script before running it? Download it first, inspect it, then run it. See Installation for pinned versions, PATH options, and building from source.
alphacode doctor # checks providers, terminal and paths, and tells you what's missing
alphacode # launch the terminal UI
No API key is required to start. The built-in free AI lane works out of the box. To use your own provider, run alphacode login (see Bring your own model). The first launch shows a short onboarding wizard (telemetry choice, model defaults, key bindings), and every step can be skipped with Esc.
Explain this project and identify the main entry points.
Find the biggest bugs, fix the highest-impact one, and run the tests.
Or run a single task non-interactively:
alphacode run "fix the failing test"
Install the AlphaCode Browser Agent extension, then:
alphacode browser setup
alphacode browser status
A healthy setup reports that the browser bridge is available and responding.
| Capability | Details |
|---|---|
| Autonomous coding agent | Natural-language tasks, multi-step planning, self-verification |
| File and code tools | Read, write, targeted edits, multi-edit, patches, code-aware search |
| Terminal | Shell execution with safety controls, background jobs, batching |
| Web | Search, fetch, scraping, HTTP flow analysis |
| Real Firefox automation | DOM, forms, tabs, frames, screenshots, downloads, existing sessions |
| Desktop automation | Windows UI Automation, macOS Accessibility API, Linux AT-SPI2 |
| Swarm Mode | Multiple cooperating agents for decomposable work |
| Persistent sessions and memory | Resume long-running work across terminal restarts |
| MCP support | Connect external tools and services |
| Skills | Reusable, specialized workflows |
| Multi-provider | Anthropic, OpenAI, Gemini, Copilot, OpenRouter, Bedrock, Azure, self-hosted, and more |
| Free AI lane | Start without an API key |
| Cross-platform | Windows, macOS, Linux |
Read and create files, apply targeted edits and patches, search repositories, trace code paths, run builds and tests, review diffs, and diagnose errors.
Combine web search and fetching with real browser automation: inspect dynamic pages, interact with web apps, handle forms and browser state, capture screenshots, and assist with QA and security testing.
The AlphaCode Browser Agent is a Firefox extension that connects AlphaCode to a real Firefox session through native messaging. AlphaCode stays the reasoning and orchestration layer, while Firefox becomes an execution environment.
flowchart LR
U["You"] --> A["AlphaCode agent"]
A --> B["browser tool"]
B --> C["Native messaging"]
C --> D["Browser Agent<br/>Firefox extension"]
D --> F["Real Firefox"]
F --> G["Page: DOM, forms,<br/>frames, UI"]
G --> D
D --> A
A --> R["Verified result"]
A plain HTTP client fetches HTML. A browser agent experiences the page the way a user does, which matters for:
Instead of pasting HTML into a chat and asking "what should I click?", you can say:
Open the application, check the login flow, go to the dashboard,
test the form validation, and report anything suspicious.
Privacy note: the agent can act inside your real browser session. Review what it is asked to do, and treat page content as untrusted input.
AlphaCode drives native applications through accessibility APIs rather than relying only on screen coordinates.
| Platform | Backend |
|---|---|
| Windows | UI Automation |
| macOS | Accessibility API |
| Linux | AT-SPI2 |
Supported actions include list_windows, snapshot, find, click, type, press, scroll, focus, screenshot, toggle, expand and collapse.
Open Calculator, calculate 123 × 456, and read the result.
AlphaCode finds the app, inspects its accessibility tree, locates the controls, enters the expression, reads the result, and reports 56088.
Large goals can be split across multiple agents working in parallel.
flowchart TD
G["Large goal"] --> P["Planner"]
P --> A["Agent A<br/>code analysis"]
P --> B["Agent B<br/>security review"]
P --> C["Agent C<br/>tests"]
P --> D["Agent D<br/>documentation"]
A --> M["Coordinator"]
B --> M
C --> M
D --> M
M --> R["Review"]
R --> F["Final result"]