$ loading_
Add enforceable safety proxies to AI coding tools to reduce destructive risks.
Copy the install command and let the AI configure it · recommended for beginners
No copy-paste install info for "mcp-augment" yet — see the docs or source repo.
Use safe_read to inspect src/auth.js, then use safe_edit to fix the login validation logic; do not delete files or reveal any environment variables or secrets.
Provides validated read and edit results and completes the intended code fix safely.
Use safe_bash to clean temporary build artifacts, but if the command includes recursive deletion, overwriting system directories, or other high-risk actions, refuse and suggest a safer alternative.
Returns results for allowed safe commands, or blocks risky commands with reasons and safer alternatives.
Review the configuration files and help me update the deployment script, but only through the safe proxy tools; if you find API keys, tokens, or passwords, never print them in plain text.
Completes config review and script update guidance while blocking, masking, or refusing secret disclosure.
Safely run Python code with AI and MCP tool integration.
Intercept and block MCP tool calls with YAML policies for safer AI agents.
Provide AI coding assistants with safe planning, task execution, and command allow-list control.
Safely let AI read, write, move, copy, and manage files.
Scan MCP server configs for injections, secrets, and dangerous commands.
Let AI coding tools safely call Codex for development and automation tasks.