---
title: Tip: 'trust this folder' in an AI coding tool can run code; review agent config and hooks like you would a build script | ctxsec
url: https://ctxsec.com/items/tip-treat-repo-agent-config-files-and-hooks-as-code-review-before-trusting-a-workspace
description: "Bitdefender's GitHub campaign shows that agent instruction files, session-start hooks and MCP configs in a repository are an execution path, not documentation."
status: "guide (based on Bitdefender Labs findings, Nov 17, 2026)"
published: 2026-11-19
---

> Bitdefender's GitHub campaign shows that agent instruction files, session-start hooks and MCP configs in a repository are an execution path, not documentation.

[← Tips & Guides](/tips) · [markdown](/items/tip-treat-repo-agent-config-files-and-hooks-as-code-review-before-trusting-a-workspace.md)

guide (based on Bitdefender Labs findings, Nov 17, 2026)

2026-11-19 11:45 PM ET

### Tip: 'trust this folder' in an AI coding tool can run code; review agent config and hooks like you would a build script

- **Open untrusted repos untrusted:** use restricted or untrusted-workspace mode, or a throwaway VM/container without your real credentials, until you've reviewed them.
- **Diff agent files in PRs:** add CODEOWNERS for AGENTS.md, CLAUDE.md, .claude/, .cursor/, .mcp.json and .vscode/ so changes get a human review.
- **Watch your own config:** keep ~/.claude, ~/.cursor, ~/.codex and similar folders under version control or file-integrity monitoring so injected hooks or MCP servers stand out.
- **Shrink token blast radius:** fine-grained, short-lived GitHub tokens without `workflow` scope unless needed.

Sources: [Bitdefender Labs: How AI agents become new targets in GitHub malware campaign (Nov 17, 2026)](https://www.bitdefender.com/en-us/blog/labs/how-ai-agents-become-new-targets-in-github-malware-campaign)
