{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreigjjzraa7qxzre4dowsyz4nk4jwiyfy4d636pq5whqxtnlvmklr7u",
    "uri": "at://did:plc:f2fwojywviapy7mx4uk5sf4d/app.bsky.feed.post/3mlq75crkyav2"
  },
  "coverImage": {
    "$type": "blob",
    "ref": {
      "$link": "bafkreifwcu4p3bf6vbr777noi2j4qjbrj7r6q4s777d6qsactwvqdnp6ba"
    },
    "mimeType": "image/png",
    "size": 256445
  },
  "path": "/ai-agents-security-for-developers-dont-let-your-agents-become-a-liability/",
  "publishedAt": "2026-05-12T16:18:15.000Z",
  "site": "https://blog.gitguardian.com",
  "tags": [
    "Artificial Intelligence"
  ],
  "textContent": "Using Cursor, GitHub Copilot, Claude Code, Codex, or another coding agent means giving software access to more than your code. It can also see the credentials available in your workspace, shell, config files, and development environment.",
  "title": "AI Agents Security for Developers: Don't Let Your Agents Become a Liability"
}