Skip to content

fix: add llms.txt #17

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
murilopmachado merged 1 commit into from
Aug 4, 2026
Merged

fix: add llms.txt #17

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
29 changes: 29 additions & 0 deletions web/llms.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,29 @@
# aigate

> OS-level sandbox for AI coding agents. Kernel-enforced file, command, and network isolation — like a Python venv, but for AI permissions.

aigate wraps Claude Code, Cursor, Copilot, Aider, and any other AI coding tool in a sandbox enforced by the operating system kernel, not the application. Application-level permission systems (`.claudeignore`, `permissions.deny`) can be bypassed; aigate uses POSIX ACLs, Linux namespaces (bubblewrap), and egress allowlists so the restriction can't be worked around from inside the tool.

## How it isolates

- **File isolation**: POSIX ACLs (Linux) / macOS ACLs deny read access to secrets, regardless of what the AI tool tries
- **Filesystem scope**: `allowed_paths` hides everything in `$HOME` except the working directory and explicitly allowed paths — non-scoped files are invisible at the kernel level, for both reads and writes
- **Process isolation**: Bubblewrap (`bwrap`) + mount namespaces on Linux; Seatbelt on macOS
- **Network isolation**: `bwrap --unshare-net` + `slirp4netns` + `iptables`/`ip6tables` restrict egress to an allowlist of domains
- **Command blocking**: deny execution of specific commands (e.g. `curl`, `wget`, `ssh`)
- **Output masking**: redact secrets (API keys, tokens) from stdout/stderr before they reach the terminal
- **Audit log + dashboard**: every run and blocked action is recorded to `~/.aigate/audit.jsonl`; `aigate serve` exposes a local web dashboard over it

## Configuration

Global config lives at `~/.aigate/config.yaml`, created by `aigate init` with sensible defaults (deny rules for `.env`, `secrets/`, `.ssh/`, `*.pem`, etc.). Projects can extend the global rules with a `.aigate.yaml` in the project root, adding `deny_read` and `allow_net` entries per project.

## Who it's for

Built for GDPR/ISO/security-regulated teams that need AI coding agents to run against real codebases without a proven-bypassable, app-level permission layer as the only safeguard. Ships as a single Go binary for Linux, macOS, and WSL.

## Contact & links

- Repository: https://github.com/AxeForging/aigate
- Website: https://axeforging.github.io/aigate/
- License: MIT
Loading